diff --git a/docs/_build/doctrees/environment.pickle b/docs/_build/doctrees/environment.pickle
index b5d2a32cc7b84de76609c5e82123a94be3160a1f..8177ee83cc892e37ca93cbc2fa2760d075755b87 100644
Binary files a/docs/_build/doctrees/environment.pickle and b/docs/_build/doctrees/environment.pickle differ
diff --git a/docs/_build/doctrees/modules.doctree b/docs/_build/doctrees/modules.doctree
deleted file mode 100644
index f4b5ab1c6821b4bd1146fc514f26bd0e9cf83af2..0000000000000000000000000000000000000000
Binary files a/docs/_build/doctrees/modules.doctree and /dev/null differ
diff --git a/docs/_build/doctrees/modules/pysegcnn.core.cli.doctree b/docs/_build/doctrees/modules/pysegcnn.core.cli.doctree
deleted file mode 100644
index e81222aa7975321050fa44856af1818779d651f1..0000000000000000000000000000000000000000
Binary files a/docs/_build/doctrees/modules/pysegcnn.core.cli.doctree and /dev/null differ
diff --git a/docs/_build/doctrees/modules/pysegcnn.core.constants.doctree b/docs/_build/doctrees/modules/pysegcnn.core.constants.doctree
deleted file mode 100644
index f7ef0115fbb77442bba1a7026cb3ebc36d90736a..0000000000000000000000000000000000000000
Binary files a/docs/_build/doctrees/modules/pysegcnn.core.constants.doctree and /dev/null differ
diff --git a/docs/_build/doctrees/modules/pysegcnn.core.dataset.doctree b/docs/_build/doctrees/modules/pysegcnn.core.dataset.doctree
deleted file mode 100644
index 750417c49169527f3062f7259bc39aa8fe1e65a0..0000000000000000000000000000000000000000
Binary files a/docs/_build/doctrees/modules/pysegcnn.core.dataset.doctree and /dev/null differ
diff --git a/docs/_build/doctrees/readme.doctree b/docs/_build/doctrees/readme.doctree
deleted file mode 100644
index a30cca3727e1fd6192fd34b4fad6b7aefd97bf08..0000000000000000000000000000000000000000
Binary files a/docs/_build/doctrees/readme.doctree and /dev/null differ
diff --git a/docs/_build/doctrees/source/api.doctree b/docs/_build/doctrees/source/api.doctree
index ddc9fbb6751ea5f248a723f91a03c3bcea5d78fe..e95d50af9bdf74e2cc1ed8fe7def65694ec5bfc3 100644
Binary files a/docs/_build/doctrees/source/api.doctree and b/docs/_build/doctrees/source/api.doctree differ
diff --git a/docs/_build/doctrees/source/core.doctree b/docs/_build/doctrees/source/core.doctree
deleted file mode 100644
index 4284a31b29f3c1dcf035cddd700c98f325eec1d3..0000000000000000000000000000000000000000
Binary files a/docs/_build/doctrees/source/core.doctree and /dev/null differ
diff --git a/docs/_build/doctrees/source/generated/pysegcnn.core.dataset.Cloud95Dataset.doctree b/docs/_build/doctrees/source/generated/pysegcnn.core.dataset.Cloud95Dataset.doctree
index 0d580be980e1cdecbb3dd9284ecf7a3905bc9ab3..5ef86132fb7c28b8990bc6e08b74291855adb746 100644
Binary files a/docs/_build/doctrees/source/generated/pysegcnn.core.dataset.Cloud95Dataset.doctree and b/docs/_build/doctrees/source/generated/pysegcnn.core.dataset.Cloud95Dataset.doctree differ
diff --git a/docs/_build/doctrees/source/generated/pysegcnn.core.dataset.ImageDataset.doctree b/docs/_build/doctrees/source/generated/pysegcnn.core.dataset.ImageDataset.doctree
index d3d8263e389faa1a9785e2f986c1a8c8b6867056..419113a28a9c3ab7364d1f88fa99c703ebdd275f 100644
Binary files a/docs/_build/doctrees/source/generated/pysegcnn.core.dataset.ImageDataset.doctree and b/docs/_build/doctrees/source/generated/pysegcnn.core.dataset.ImageDataset.doctree differ
diff --git a/docs/_build/doctrees/source/generated/pysegcnn.core.dataset.SparcsDataset.doctree b/docs/_build/doctrees/source/generated/pysegcnn.core.dataset.SparcsDataset.doctree
index 09ffadd51e5d992e099585385ba906deb751727c..20aad787c0af6cccc82308a42b2f2995910e626d 100644
Binary files a/docs/_build/doctrees/source/generated/pysegcnn.core.dataset.SparcsDataset.doctree and b/docs/_build/doctrees/source/generated/pysegcnn.core.dataset.SparcsDataset.doctree differ
diff --git a/docs/_build/doctrees/source/generated/pysegcnn.core.layers.Block.doctree b/docs/_build/doctrees/source/generated/pysegcnn.core.layers.Block.doctree
index cea7dcb2ebd3e2955e4740ba5e433b8a9272b9f6..23c8160fee9d36e17b17ef1ba2e82de2fad14887 100644
Binary files a/docs/_build/doctrees/source/generated/pysegcnn.core.layers.Block.doctree and b/docs/_build/doctrees/source/generated/pysegcnn.core.layers.Block.doctree differ
diff --git a/docs/_build/doctrees/source/generated/pysegcnn.core.layers.Conv2dSame.doctree b/docs/_build/doctrees/source/generated/pysegcnn.core.layers.Conv2dSame.doctree
index aa7807531296decd72e116b6c2622581b031ee6e..69c0fd72092fcdc2070369b78e183ae42e5e3190 100644
Binary files a/docs/_build/doctrees/source/generated/pysegcnn.core.layers.Conv2dSame.doctree and b/docs/_build/doctrees/source/generated/pysegcnn.core.layers.Conv2dSame.doctree differ
diff --git a/docs/_build/doctrees/source/generated/pysegcnn.core.layers.ConvBnReluMaxPool.doctree b/docs/_build/doctrees/source/generated/pysegcnn.core.layers.ConvBnReluMaxPool.doctree
index 541b14d3a65f0b1f66766b2b1f461192b96948d2..8137ca9dd91d89ad824f5f5fd3171fa0b892eb5e 100644
Binary files a/docs/_build/doctrees/source/generated/pysegcnn.core.layers.ConvBnReluMaxPool.doctree and b/docs/_build/doctrees/source/generated/pysegcnn.core.layers.ConvBnReluMaxPool.doctree differ
diff --git a/docs/_build/doctrees/source/generated/pysegcnn.core.layers.ConvBnReluMaxUnpool.doctree b/docs/_build/doctrees/source/generated/pysegcnn.core.layers.ConvBnReluMaxUnpool.doctree
index bc1ac7f948be56eff1ce5bb5e6c0c4062c16a12a..bc7a303b51b774ecc224b78ededae4354a6ba137 100644
Binary files a/docs/_build/doctrees/source/generated/pysegcnn.core.layers.ConvBnReluMaxUnpool.doctree and b/docs/_build/doctrees/source/generated/pysegcnn.core.layers.ConvBnReluMaxUnpool.doctree differ
diff --git a/docs/_build/doctrees/source/generated/pysegcnn.core.layers.ConvBnReluUpsample.doctree b/docs/_build/doctrees/source/generated/pysegcnn.core.layers.ConvBnReluUpsample.doctree
index 138065b0fe4285d469ccde5dac5d702365626cd0..7ae0b90e42a8438aeb0b968bbf44dac505134158 100644
Binary files a/docs/_build/doctrees/source/generated/pysegcnn.core.layers.ConvBnReluUpsample.doctree and b/docs/_build/doctrees/source/generated/pysegcnn.core.layers.ConvBnReluUpsample.doctree differ
diff --git a/docs/_build/doctrees/source/generated/pysegcnn.core.layers.Decoder.doctree b/docs/_build/doctrees/source/generated/pysegcnn.core.layers.Decoder.doctree
index f42921f7fe87e17d8af1dba21fcc9d66bf68e036..c5db12d335ea6ef9b6e03c97981ebcc795e9b949 100644
Binary files a/docs/_build/doctrees/source/generated/pysegcnn.core.layers.Decoder.doctree and b/docs/_build/doctrees/source/generated/pysegcnn.core.layers.Decoder.doctree differ
diff --git a/docs/_build/doctrees/source/generated/pysegcnn.core.layers.DecoderBlock.doctree b/docs/_build/doctrees/source/generated/pysegcnn.core.layers.DecoderBlock.doctree
index 050a01236c62ddc6683a68129434b660471a5332..951a1fdb5678105b9674507c191f4ea840e50f34 100644
Binary files a/docs/_build/doctrees/source/generated/pysegcnn.core.layers.DecoderBlock.doctree and b/docs/_build/doctrees/source/generated/pysegcnn.core.layers.DecoderBlock.doctree differ
diff --git a/docs/_build/doctrees/source/generated/pysegcnn.core.layers.Encoder.doctree b/docs/_build/doctrees/source/generated/pysegcnn.core.layers.Encoder.doctree
index 443e2179a37b744b3483d9770eaed40988d9e567..3c96ed7f91ef33f310ac27623fbe2e44ce193639 100644
Binary files a/docs/_build/doctrees/source/generated/pysegcnn.core.layers.Encoder.doctree and b/docs/_build/doctrees/source/generated/pysegcnn.core.layers.Encoder.doctree differ
diff --git a/docs/_build/doctrees/source/generated/pysegcnn.core.layers.EncoderBlock.doctree b/docs/_build/doctrees/source/generated/pysegcnn.core.layers.EncoderBlock.doctree
index 140acc3978a3b14f87cecee09fb329a27f7b17ba..c995fe994747a627df70b5eae6aaef20da4f50fa 100644
Binary files a/docs/_build/doctrees/source/generated/pysegcnn.core.layers.EncoderBlock.doctree and b/docs/_build/doctrees/source/generated/pysegcnn.core.layers.EncoderBlock.doctree differ
diff --git a/docs/_build/doctrees/source/generated/pysegcnn.core.layers.doctree b/docs/_build/doctrees/source/generated/pysegcnn.core.layers.doctree
deleted file mode 100644
index 5cee477b5e57fbafaed6fdddff78b47cc155da38..0000000000000000000000000000000000000000
Binary files a/docs/_build/doctrees/source/generated/pysegcnn.core.layers.doctree and /dev/null differ
diff --git a/docs/_build/doctrees/source/generated/pysegcnn.core.models.Network.doctree b/docs/_build/doctrees/source/generated/pysegcnn.core.models.Network.doctree
index b897c77c0e8f679d5d321aca5a887450fbbb06fe..3b66c63724c97ade22e24070a3b643f8c9103532 100644
Binary files a/docs/_build/doctrees/source/generated/pysegcnn.core.models.Network.doctree and b/docs/_build/doctrees/source/generated/pysegcnn.core.models.Network.doctree differ
diff --git a/docs/_build/doctrees/source/generated/pysegcnn.core.models.UNet.doctree b/docs/_build/doctrees/source/generated/pysegcnn.core.models.UNet.doctree
index b2dbf831d1a9438737d3807c76722aa3d44862eb..b89ad75eb33b76b2c53d7e87facdcb23b1b23395 100644
Binary files a/docs/_build/doctrees/source/generated/pysegcnn.core.models.UNet.doctree and b/docs/_build/doctrees/source/generated/pysegcnn.core.models.UNet.doctree differ
diff --git a/docs/_build/doctrees/source/generated/pysegcnn.core.models.doctree b/docs/_build/doctrees/source/generated/pysegcnn.core.models.doctree
deleted file mode 100644
index c0daf6db0f35aa1799069dd78e475da38cf05b35..0000000000000000000000000000000000000000
Binary files a/docs/_build/doctrees/source/generated/pysegcnn.core.models.doctree and /dev/null differ
diff --git a/docs/_build/doctrees/source/install.doctree b/docs/_build/doctrees/source/install.doctree
deleted file mode 100644
index 9b01515b3c55b59c5170d8c51e6ff4ef1fc4040a..0000000000000000000000000000000000000000
Binary files a/docs/_build/doctrees/source/install.doctree and /dev/null differ
diff --git a/docs/_build/doctrees/source/main.doctree b/docs/_build/doctrees/source/main.doctree
deleted file mode 100644
index 765811a4dbdd68a8f38b5c5f4f6ca1634cbada14..0000000000000000000000000000000000000000
Binary files a/docs/_build/doctrees/source/main.doctree and /dev/null differ
diff --git a/docs/_build/doctrees/source/modules.doctree b/docs/_build/doctrees/source/modules.doctree
deleted file mode 100644
index df0a22b202133625edfbfdd0f85c30b6070565e8..0000000000000000000000000000000000000000
Binary files a/docs/_build/doctrees/source/modules.doctree and /dev/null differ
diff --git a/docs/_build/doctrees/source/preprocessing.doctree b/docs/_build/doctrees/source/preprocessing.doctree
deleted file mode 100644
index 9447189fbe4a1974df0cdac3501add9a5d4e10f4..0000000000000000000000000000000000000000
Binary files a/docs/_build/doctrees/source/preprocessing.doctree and /dev/null differ
diff --git a/docs/_build/doctrees/source/readme.doctree b/docs/_build/doctrees/source/readme.doctree
deleted file mode 100644
index 1fda9e038780a91cbb0ab13fa6d2d65f95a88802..0000000000000000000000000000000000000000
Binary files a/docs/_build/doctrees/source/readme.doctree and /dev/null differ
diff --git a/docs/_build/html/_modules/core/cli.html b/docs/_build/html/_modules/core/cli.html
deleted file mode 100644
index d73d939362fb8c78c8d6ea841ac865401f2d21aa..0000000000000000000000000000000000000000
--- a/docs/_build/html/_modules/core/cli.html
+++ /dev/null
@@ -1,264 +0,0 @@
-
-
-<!DOCTYPE html>
-<html class="writer-html5" lang="en" >
-<head>
- <meta charset="utf-8">
-
- <meta name="viewport" content="width=device-width, initial-scale=1.0">
-
- <title>core.cli — PySegCNN documentation</title>
-
-
-
- <link rel="stylesheet" href="../../_static/css/theme.css" type="text/css" />
- <link rel="stylesheet" href="../../_static/pygments.css" type="text/css" />
-
-
-
-
-
-
-
- <!--[if lt IE 9]>
- <script src="../../_static/js/html5shiv.min.js"></script>
- <![endif]-->
-
-
- <script type="text/javascript" id="documentation_options" data-url_root="../../" src="../../_static/documentation_options.js"></script>
- <script src="../../_static/jquery.js"></script>
- <script src="../../_static/underscore.js"></script>
- <script src="../../_static/doctools.js"></script>
- <script src="../../_static/language_data.js"></script>
-
- <script type="text/javascript" src="../../_static/js/theme.js"></script>
-
-
- <link rel="index" title="Index" href="../../genindex.html" />
- <link rel="search" title="Search" href="../../search.html" />
-</head>
-
-<body class="wy-body-for-nav">
-
-
- <div class="wy-grid-for-nav">
-
- <nav data-toggle="wy-nav-shift" class="wy-nav-side">
- <div class="wy-side-scroll">
- <div class="wy-side-nav-search" >
-
-
-
- <a href="../../index.html" class="icon icon-home" alt="Documentation Home"> PySegCNN
-
-
-
- </a>
-
-
-
-
-
-
-
-<div role="search">
- <form id="rtd-search-form" class="wy-form" action="../../search.html" method="get">
- <input type="text" name="q" placeholder="Search docs" />
- <input type="hidden" name="check_keywords" value="yes" />
- <input type="hidden" name="area" value="default" />
- </form>
-</div>
-
-
- </div>
-
-
- <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
-
-
-
-
-
-
- <ul>
-<li class="toctree-l1"><a class="reference internal" href="../../source/requirements.html">Requirements</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../../source/installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../../source/api.html">API Reference</a></li>
-</ul>
-
-
-
- </div>
-
- </div>
- </nav>
-
- <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
-
-
- <nav class="wy-nav-top" aria-label="top navigation">
-
- <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
- <a href="../../index.html">PySegCNN</a>
-
- </nav>
-
-
- <div class="wy-nav-content">
-
- <div class="rst-content">
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-<div role="navigation" aria-label="breadcrumbs navigation">
-
- <ul class="wy-breadcrumbs">
-
- <li><a href="../../index.html" class="icon icon-home"></a> »</li>
-
- <li><a href="../index.html">Module code</a> »</li>
-
- <li>core.cli</li>
-
-
- <li class="wy-breadcrumbs-aside">
-
- </li>
-
- </ul>
-
-
- <hr/>
-</div>
- <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
- <div itemprop="articleBody">
-
- <h1>Source code for core.cli</h1><div class="highlight"><pre>
-<span></span><span class="sd">"""Command line interface parsers.</span>
-
-<span class="sd">License</span>
-<span class="sd">-------</span>
-
-<span class="sd"> Copyright (c) 2020 Daniel Frisinghelli</span>
-
-<span class="sd"> This source code is licensed under the GNU General Public License v3.</span>
-
-<span class="sd"> See the LICENSE file in the repository's root directory.</span>
-
-<span class="sd">"""</span>
-
-<span class="c1"># !/usr/bin/env python</span>
-<span class="c1"># -*- coding: utf-8 -*-</span>
-
-<span class="c1"># builtins</span>
-<span class="kn">import</span> <span class="nn">pathlib</span>
-<span class="kn">import</span> <span class="nn">argparse</span>
-
-
-<span class="c1"># epilogue to display at the end of each parser</span>
-<span class="n">EPILOGUE</span> <span class="o">=</span> <span class="s1">'Author: Daniel Frisinghelli, daniel.frisinghelli@gmail.com'</span>
-
-
-<div class="viewcode-block" id="structure_parser"><a class="viewcode-back" href="../../source/core.html#core.cli.structure_parser">[docs]</a><span class="k">def</span> <span class="nf">structure_parser</span><span class="p">():</span>
- <span class="sd">"""Command line argument parser to standardize dataset structure.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
- <span class="n">parser</span> <span class="o">=</span> <span class="n">argparse</span><span class="o">.</span><span class="n">ArgumentParser</span><span class="p">(</span>
- <span class="n">description</span><span class="o">=</span><span class="s1">'Standardize the dataset directory structure.'</span><span class="p">,</span>
- <span class="n">epilog</span><span class="o">=</span><span class="n">EPILOGUE</span><span class="p">,</span>
- <span class="n">formatter_class</span><span class="o">=</span><span class="k">lambda</span> <span class="n">prog</span><span class="p">:</span> <span class="n">argparse</span><span class="o">.</span><span class="n">RawDescriptionHelpFormatter</span><span class="p">(</span>
- <span class="n">prog</span><span class="p">,</span> <span class="n">max_help_position</span><span class="o">=</span><span class="mi">50</span><span class="p">,</span> <span class="n">indent_increment</span><span class="o">=</span><span class="mi">2</span><span class="p">))</span>
-
- <span class="c1"># positional arguments</span>
-
- <span class="c1"># positional argument: path to the archive</span>
- <span class="n">parser</span><span class="o">.</span><span class="n">add_argument</span><span class="p">(</span><span class="s1">'archive'</span><span class="p">,</span> <span class="nb">type</span><span class="o">=</span><span class="n">pathlib</span><span class="o">.</span><span class="n">Path</span><span class="p">,</span>
- <span class="n">help</span><span class="o">=</span><span class="s1">'Path to the dataset archive.'</span><span class="p">)</span>
-
- <span class="c1"># positional argument: path to extract and restructure the dataset</span>
- <span class="n">parser</span><span class="o">.</span><span class="n">add_argument</span><span class="p">(</span><span class="s1">'target'</span><span class="p">,</span> <span class="nb">type</span><span class="o">=</span><span class="n">pathlib</span><span class="o">.</span><span class="n">Path</span><span class="p">,</span>
- <span class="n">help</span><span class="o">=</span><span class="s1">'Path to save standardized dataset structure.'</span><span class="p">)</span>
-
- <span class="c1"># optional arguments</span>
-
- <span class="c1"># default values</span>
- <span class="n">default</span> <span class="o">=</span> <span class="s1">'(default: </span><span class="si">%(default)s</span><span class="s1">)'</span>
-
- <span class="c1"># optional argument: whether to overwrite existing files</span>
- <span class="n">parser</span><span class="o">.</span><span class="n">add_argument</span><span class="p">(</span><span class="s1">'-o'</span><span class="p">,</span> <span class="s1">'--overwrite'</span><span class="p">,</span> <span class="nb">type</span><span class="o">=</span><span class="nb">bool</span><span class="p">,</span>
- <span class="n">help</span><span class="o">=</span><span class="s1">'Overwrite files </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">default</span><span class="p">),</span>
- <span class="n">default</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">nargs</span><span class="o">=</span><span class="s1">'?'</span><span class="p">,</span> <span class="n">const</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span> <span class="n">metavar</span><span class="o">=</span><span class="s1">''</span><span class="p">)</span>
-
- <span class="c1"># optional argument: whether to copy or move extracted files</span>
- <span class="n">parser</span><span class="o">.</span><span class="n">add_argument</span><span class="p">(</span><span class="s1">'-r'</span><span class="p">,</span> <span class="s1">'--remove'</span><span class="p">,</span> <span class="nb">type</span><span class="o">=</span><span class="nb">bool</span><span class="p">,</span>
- <span class="n">help</span><span class="o">=</span><span class="s1">'Remove original dataset </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">default</span><span class="p">),</span>
- <span class="n">default</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span> <span class="n">nargs</span><span class="o">=</span><span class="s1">'?'</span><span class="p">,</span> <span class="n">const</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span> <span class="n">metavar</span><span class="o">=</span><span class="s1">''</span><span class="p">)</span>
-
- <span class="k">return</span> <span class="n">parser</span></div>
-</pre></div>
-
- </div>
-
- </div>
- <footer>
-
-
- <hr/>
-
- <div role="contentinfo">
- <p>
-
- © Copyright 2020, Daniel Frisinghelli
-
- </p>
- </div>
-
-
-
- Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a
-
- <a href="https://github.com/rtfd/sphinx_rtd_theme">theme</a>
-
- provided by <a href="https://readthedocs.org">Read the Docs</a>.
-
-</footer>
-
- </div>
- </div>
-
- </section>
-
- </div>
-
-
- <script type="text/javascript">
- jQuery(function () {
- SphinxRtdTheme.Navigation.enable(true);
- });
- </script>
-
-
-
-
-
-
-</body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/_modules/core/constants.html b/docs/_build/html/_modules/core/constants.html
deleted file mode 100644
index 046478e122c847eff12af3f24609a186efd49625..0000000000000000000000000000000000000000
--- a/docs/_build/html/_modules/core/constants.html
+++ /dev/null
@@ -1,314 +0,0 @@
-
-
-<!DOCTYPE html>
-<html class="writer-html5" lang="en" >
-<head>
- <meta charset="utf-8">
-
- <meta name="viewport" content="width=device-width, initial-scale=1.0">
-
- <title>core.constants — PySegCNN documentation</title>
-
-
-
- <link rel="stylesheet" href="../../_static/css/theme.css" type="text/css" />
- <link rel="stylesheet" href="../../_static/pygments.css" type="text/css" />
-
-
-
-
-
-
-
- <!--[if lt IE 9]>
- <script src="../../_static/js/html5shiv.min.js"></script>
- <![endif]-->
-
-
- <script type="text/javascript" id="documentation_options" data-url_root="../../" src="../../_static/documentation_options.js"></script>
- <script src="../../_static/jquery.js"></script>
- <script src="../../_static/underscore.js"></script>
- <script src="../../_static/doctools.js"></script>
- <script src="../../_static/language_data.js"></script>
-
- <script type="text/javascript" src="../../_static/js/theme.js"></script>
-
-
- <link rel="index" title="Index" href="../../genindex.html" />
- <link rel="search" title="Search" href="../../search.html" />
-</head>
-
-<body class="wy-body-for-nav">
-
-
- <div class="wy-grid-for-nav">
-
- <nav data-toggle="wy-nav-shift" class="wy-nav-side">
- <div class="wy-side-scroll">
- <div class="wy-side-nav-search" >
-
-
-
- <a href="../../index.html" class="icon icon-home" alt="Documentation Home"> PySegCNN
-
-
-
- </a>
-
-
-
-
-
-
-
-<div role="search">
- <form id="rtd-search-form" class="wy-form" action="../../search.html" method="get">
- <input type="text" name="q" placeholder="Search docs" />
- <input type="hidden" name="check_keywords" value="yes" />
- <input type="hidden" name="area" value="default" />
- </form>
-</div>
-
-
- </div>
-
-
- <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
-
-
-
-
-
-
- <ul>
-<li class="toctree-l1"><a class="reference internal" href="../../source/requirements.html">Requirements</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../../source/installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../../source/api.html">API Reference</a></li>
-</ul>
-
-
-
- </div>
-
- </div>
- </nav>
-
- <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
-
-
- <nav class="wy-nav-top" aria-label="top navigation">
-
- <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
- <a href="../../index.html">PySegCNN</a>
-
- </nav>
-
-
- <div class="wy-nav-content">
-
- <div class="rst-content">
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-<div role="navigation" aria-label="breadcrumbs navigation">
-
- <ul class="wy-breadcrumbs">
-
- <li><a href="../../index.html" class="icon icon-home"></a> »</li>
-
- <li><a href="../index.html">Module code</a> »</li>
-
- <li>core.constants</li>
-
-
- <li class="wy-breadcrumbs-aside">
-
- </li>
-
- </ul>
-
-
- <hr/>
-</div>
- <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
- <div itemprop="articleBody">
-
- <h1>Source code for core.constants</h1><div class="highlight"><pre>
-<span></span><span class="sd">"""A collection of constant values.</span>
-
-<span class="sd">License</span>
-<span class="sd">-------</span>
-
-<span class="sd"> Copyright (c) 2020 Daniel Frisinghelli</span>
-
-<span class="sd"> This source code is licensed under the GNU General Public License v3.</span>
-
-<span class="sd"> See the LICENSE file in the repository's root directory.</span>
-
-<span class="sd">"""</span>
-
-<span class="c1"># !/usr/bin/env python</span>
-<span class="c1"># -*- coding: utf-8 -*-</span>
-
-<span class="c1"># builtins</span>
-<span class="kn">import</span> <span class="nn">enum</span>
-
-
-<span class="c1"># Landsat 8 bands</span>
-<div class="viewcode-block" id="Landsat8"><a class="viewcode-back" href="../../source/core.html#core.constants.Landsat8">[docs]</a><span class="k">class</span> <span class="nc">Landsat8</span><span class="p">(</span><span class="n">enum</span><span class="o">.</span><span class="n">Enum</span><span class="p">):</span>
- <span class="sd">"""The spectral bands of the Landsat 8 sensors.</span>
-
-<span class="sd"> sensors:</span>
-<span class="sd"> - Operational Land Imager (OLI), (bands 1-9)</span>
-<span class="sd"> - Thermal Infrared Sensor (TIRS), (bands 10, 11)</span>
-
-<span class="sd"> """</span>
-
- <span class="n">violet</span> <span class="o">=</span> <span class="mi">1</span>
- <span class="n">blue</span> <span class="o">=</span> <span class="mi">2</span>
- <span class="n">green</span> <span class="o">=</span> <span class="mi">3</span>
- <span class="n">red</span> <span class="o">=</span> <span class="mi">4</span>
- <span class="n">nir</span> <span class="o">=</span> <span class="mi">5</span>
- <span class="n">swir1</span> <span class="o">=</span> <span class="mi">6</span>
- <span class="n">swir2</span> <span class="o">=</span> <span class="mi">7</span>
- <span class="n">pan</span> <span class="o">=</span> <span class="mi">8</span>
- <span class="n">cirrus</span> <span class="o">=</span> <span class="mi">9</span>
- <span class="n">tir1</span> <span class="o">=</span> <span class="mi">10</span></div>
- <span class="c1"># tir2 = 11</span>
-
-
-<span class="c1"># Sentinel 2 bands</span>
-<div class="viewcode-block" id="Sentinel2"><a class="viewcode-back" href="../../source/core.html#core.constants.Sentinel2">[docs]</a><span class="k">class</span> <span class="nc">Sentinel2</span><span class="p">(</span><span class="n">enum</span><span class="o">.</span><span class="n">Enum</span><span class="p">):</span>
- <span class="sd">"""The spectral bands of the Sentinel-2 MultiSpectral Instrument (MSI)."""</span>
-
- <span class="n">aerosol</span> <span class="o">=</span> <span class="mi">1</span>
- <span class="n">blue</span> <span class="o">=</span> <span class="mi">2</span>
- <span class="n">green</span> <span class="o">=</span> <span class="mi">3</span>
- <span class="n">red</span> <span class="o">=</span> <span class="mi">4</span>
- <span class="n">vnir1</span> <span class="o">=</span> <span class="mi">5</span>
- <span class="n">vnir2</span> <span class="o">=</span> <span class="mi">6</span>
- <span class="n">vnir3</span> <span class="o">=</span> <span class="mi">7</span>
- <span class="n">nir</span> <span class="o">=</span> <span class="mi">8</span>
- <span class="n">nnir</span> <span class="o">=</span> <span class="s1">'8A'</span>
- <span class="n">vapor</span> <span class="o">=</span> <span class="mi">9</span>
- <span class="n">cirrus</span> <span class="o">=</span> <span class="mi">10</span>
- <span class="n">swir1</span> <span class="o">=</span> <span class="mi">11</span>
- <span class="n">swir2</span> <span class="o">=</span> <span class="mi">12</span></div>
-
-
-<span class="c1"># generic class label enumeration class</span>
-<div class="viewcode-block" id="Label"><a class="viewcode-back" href="../../source/core.html#core.constants.Label">[docs]</a><span class="k">class</span> <span class="nc">Label</span><span class="p">(</span><span class="n">enum</span><span class="o">.</span><span class="n">Enum</span><span class="p">):</span>
- <span class="sd">"""Generic enumeration for class labels."""</span>
-
- <span class="nd">@property</span>
- <span class="k">def</span> <span class="nf">id</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Return the value of a class in the ground truth."""</span>
- <span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">value</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span>
-
- <span class="nd">@property</span>
- <span class="k">def</span> <span class="nf">color</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Return the color to plot a class."""</span>
- <span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">value</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span></div>
-
-
-<span class="c1"># labels of the Sparcs dataset</span>
-<div class="viewcode-block" id="SparcsLabels"><a class="viewcode-back" href="../../source/core.html#core.constants.SparcsLabels">[docs]</a><span class="k">class</span> <span class="nc">SparcsLabels</span><span class="p">(</span><span class="n">Label</span><span class="p">):</span>
- <span class="sd">"""Class labels of the `Sparcs`_ dataset.</span>
-
-<span class="sd"> .. _Sparcs:</span>
-<span class="sd"> https://www.usgs.gov/land-resources/nli/landsat/spatial-procedures-automated-removal-cloud-and-shadow-sparcs-validation</span>
-
-<span class="sd"> """</span>
-
- <span class="n">Shadow</span> <span class="o">=</span> <span class="mi">0</span><span class="p">,</span> <span class="s1">'grey'</span>
- <span class="n">Shadow_over_water</span> <span class="o">=</span> <span class="mi">1</span><span class="p">,</span> <span class="s1">'darkblue'</span>
- <span class="n">Water</span> <span class="o">=</span> <span class="mi">2</span><span class="p">,</span> <span class="s1">'blue'</span>
- <span class="n">Snow</span> <span class="o">=</span> <span class="mi">3</span><span class="p">,</span> <span class="s1">'lightblue'</span>
- <span class="n">Land</span> <span class="o">=</span> <span class="mi">4</span><span class="p">,</span> <span class="s1">'sienna'</span>
- <span class="n">Cloud</span> <span class="o">=</span> <span class="mi">5</span><span class="p">,</span> <span class="s1">'white'</span>
- <span class="n">Flooded</span> <span class="o">=</span> <span class="mi">6</span><span class="p">,</span> <span class="s1">'yellow'</span></div>
-
-
-<span class="c1"># labels of the Cloud95 dataset</span>
-<div class="viewcode-block" id="Cloud95Labels"><a class="viewcode-back" href="../../source/core.html#core.constants.Cloud95Labels">[docs]</a><span class="k">class</span> <span class="nc">Cloud95Labels</span><span class="p">(</span><span class="n">Label</span><span class="p">):</span>
- <span class="sd">"""Class labels of the `Cloud-95`_ dataset.</span>
-
-<span class="sd"> .. _Cloud-95:</span>
-<span class="sd"> https://github.com/SorourMo/95-Cloud-An-Extension-to-38-Cloud-Dataset</span>
-
-<span class="sd"> """</span>
-
- <span class="n">Clear</span> <span class="o">=</span> <span class="mi">0</span><span class="p">,</span> <span class="s1">'skyblue'</span>
- <span class="n">Cloud</span> <span class="o">=</span> <span class="mi">1</span><span class="p">,</span> <span class="s1">'white'</span></div>
-
-
-<span class="c1"># labels of the ProSnow dataset</span>
-<div class="viewcode-block" id="ProSnowLabels"><a class="viewcode-back" href="../../source/core.html#core.constants.ProSnowLabels">[docs]</a><span class="k">class</span> <span class="nc">ProSnowLabels</span><span class="p">(</span><span class="n">Label</span><span class="p">):</span>
- <span class="sd">"""Class labels of the ProSnow datasets."""</span>
-
- <span class="n">Cloud</span> <span class="o">=</span> <span class="mi">0</span><span class="p">,</span> <span class="s1">'white'</span>
- <span class="n">Snow</span> <span class="o">=</span> <span class="mi">1</span><span class="p">,</span> <span class="s1">'lightblue'</span>
- <span class="n">Snow_free</span> <span class="o">=</span> <span class="mi">2</span><span class="p">,</span> <span class="s1">'sienna'</span></div>
-</pre></div>
-
- </div>
-
- </div>
- <footer>
-
-
- <hr/>
-
- <div role="contentinfo">
- <p>
-
- © Copyright 2020, Daniel Frisinghelli
-
- </p>
- </div>
-
-
-
- Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a
-
- <a href="https://github.com/rtfd/sphinx_rtd_theme">theme</a>
-
- provided by <a href="https://readthedocs.org">Read the Docs</a>.
-
-</footer>
-
- </div>
- </div>
-
- </section>
-
- </div>
-
-
- <script type="text/javascript">
- jQuery(function () {
- SphinxRtdTheme.Navigation.enable(true);
- });
- </script>
-
-
-
-
-
-
-</body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/_modules/core/dataset.html b/docs/_build/html/_modules/core/dataset.html
deleted file mode 100644
index 8b56cc4a6d83fc94d4da3de1595f26d220538245..0000000000000000000000000000000000000000
--- a/docs/_build/html/_modules/core/dataset.html
+++ /dev/null
@@ -1,1594 +0,0 @@
-
-
-<!DOCTYPE html>
-<html class="writer-html5" lang="en" >
-<head>
- <meta charset="utf-8">
-
- <meta name="viewport" content="width=device-width, initial-scale=1.0">
-
- <title>core.dataset — PySegCNN documentation</title>
-
-
-
- <link rel="stylesheet" href="../../_static/css/theme.css" type="text/css" />
- <link rel="stylesheet" href="../../_static/pygments.css" type="text/css" />
-
-
-
-
-
-
-
- <!--[if lt IE 9]>
- <script src="../../_static/js/html5shiv.min.js"></script>
- <![endif]-->
-
-
- <script type="text/javascript" id="documentation_options" data-url_root="../../" src="../../_static/documentation_options.js"></script>
- <script src="../../_static/jquery.js"></script>
- <script src="../../_static/underscore.js"></script>
- <script src="../../_static/doctools.js"></script>
- <script src="../../_static/language_data.js"></script>
-
- <script type="text/javascript" src="../../_static/js/theme.js"></script>
-
-
- <link rel="index" title="Index" href="../../genindex.html" />
- <link rel="search" title="Search" href="../../search.html" />
-</head>
-
-<body class="wy-body-for-nav">
-
-
- <div class="wy-grid-for-nav">
-
- <nav data-toggle="wy-nav-shift" class="wy-nav-side">
- <div class="wy-side-scroll">
- <div class="wy-side-nav-search" >
-
-
-
- <a href="../../index.html" class="icon icon-home" alt="Documentation Home"> PySegCNN
-
-
-
- </a>
-
-
-
-
-
-
-
-<div role="search">
- <form id="rtd-search-form" class="wy-form" action="../../search.html" method="get">
- <input type="text" name="q" placeholder="Search docs" />
- <input type="hidden" name="check_keywords" value="yes" />
- <input type="hidden" name="area" value="default" />
- </form>
-</div>
-
-
- </div>
-
-
- <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
-
-
-
-
-
-
- <ul>
-<li class="toctree-l1"><a class="reference internal" href="../../source/requirements.html">Requirements</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../../source/installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../../source/api.html">API Reference</a></li>
-</ul>
-
-
-
- </div>
-
- </div>
- </nav>
-
- <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
-
-
- <nav class="wy-nav-top" aria-label="top navigation">
-
- <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
- <a href="../../index.html">PySegCNN</a>
-
- </nav>
-
-
- <div class="wy-nav-content">
-
- <div class="rst-content">
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-<div role="navigation" aria-label="breadcrumbs navigation">
-
- <ul class="wy-breadcrumbs">
-
- <li><a href="../../index.html" class="icon icon-home"></a> »</li>
-
- <li><a href="../index.html">Module code</a> »</li>
-
- <li>core.dataset</li>
-
-
- <li class="wy-breadcrumbs-aside">
-
- </li>
-
- </ul>
-
-
- <hr/>
-</div>
- <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
- <div itemprop="articleBody">
-
- <h1>Source code for core.dataset</h1><div class="highlight"><pre>
-<span></span><span class="sd">"""Custom dataset classes compliant to the PyTorch standard.</span>
-
-<span class="sd">Each custom dataset should inherit from torch.utils.data.Dataset to benefit</span>
-<span class="sd">from the torch.utils.data.DataLoader class, which implements helpful utilities</span>
-<span class="sd">during model training.</span>
-
-<span class="sd">For any kind of image-like dataset, inherit the ImageDataset class to create</span>
-<span class="sd">your custom dataset.</span>
-
-<span class="sd">License</span>
-<span class="sd">-------</span>
-
-<span class="sd"> Copyright (c) 2020 Daniel Frisinghelli</span>
-
-<span class="sd"> This source code is licensed under the GNU General Public License v3.</span>
-
-<span class="sd"> See the LICENSE file in the repository's root directory.</span>
-
-<span class="sd">"""</span>
-
-<span class="c1"># !/usr/bin/env python</span>
-<span class="c1"># -*- coding: utf-8 -*-</span>
-
-<span class="c1"># builtins</span>
-<span class="kn">import</span> <span class="nn">os</span>
-<span class="kn">import</span> <span class="nn">re</span>
-<span class="kn">import</span> <span class="nn">csv</span>
-<span class="kn">import</span> <span class="nn">enum</span>
-<span class="kn">import</span> <span class="nn">itertools</span>
-<span class="kn">import</span> <span class="nn">logging</span>
-
-<span class="c1"># externals</span>
-<span class="kn">import</span> <span class="nn">numpy</span> <span class="k">as</span> <span class="nn">np</span>
-<span class="kn">import</span> <span class="nn">torch</span>
-<span class="kn">from</span> <span class="nn">torch.utils.data</span> <span class="kn">import</span> <span class="n">Dataset</span>
-
-<span class="c1"># locals</span>
-<span class="kn">from</span> <span class="nn">pysegcnn.core.constants</span> <span class="kn">import</span> <span class="p">(</span><span class="n">Landsat8</span><span class="p">,</span> <span class="n">Sentinel2</span><span class="p">,</span> <span class="n">Label</span><span class="p">,</span> <span class="n">SparcsLabels</span><span class="p">,</span>
- <span class="n">Cloud95Labels</span><span class="p">,</span> <span class="n">ProSnowLabels</span><span class="p">)</span>
-<span class="kn">from</span> <span class="nn">pysegcnn.core.utils</span> <span class="kn">import</span> <span class="p">(</span><span class="n">img2np</span><span class="p">,</span> <span class="n">is_divisible</span><span class="p">,</span> <span class="n">tile_topleft_corner</span><span class="p">,</span>
- <span class="n">parse_landsat_scene</span><span class="p">,</span> <span class="n">parse_sentinel2_scene</span><span class="p">)</span>
-
-<span class="c1"># module level logger</span>
-<span class="n">LOGGER</span> <span class="o">=</span> <span class="n">logging</span><span class="o">.</span><span class="n">getLogger</span><span class="p">(</span><span class="vm">__name__</span><span class="p">)</span>
-
-
-<div class="viewcode-block" id="ImageDataset"><a class="viewcode-back" href="../../source/core.html#core.dataset.ImageDataset">[docs]</a><span class="k">class</span> <span class="nc">ImageDataset</span><span class="p">(</span><span class="n">Dataset</span><span class="p">):</span>
- <span class="sa">r</span><span class="sd">"""Base class for multispectral image data.</span>
-
-<span class="sd"> Inheriting from `torch.utils.data.Dataset` to be compliant to the PyTorch</span>
-<span class="sd"> standard. Furthermore, using instances of `torch.utils.data.Dataset`</span>
-<span class="sd"> enables the use of the handy `torch.utils.data.DataLoader` class during</span>
-<span class="sd"> model training.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> root_dir : `str`</span>
-<span class="sd"> The root directory, path to the dataset.</span>
-<span class="sd"> use_bands : `list` [`str`], optional</span>
-<span class="sd"> A list of the spectral bands to use. The default is [].</span>
-<span class="sd"> tile_size : `int` or `None`, optional</span>
-<span class="sd"> The size of the tiles. If not `None`, each scene is divided into square</span>
-<span class="sd"> tiles of shape (tile_size, tile_size). The default is None.</span>
-<span class="sd"> pad : `bool`, optional</span>
-<span class="sd"> Whether to center pad the input image. Set ``pad`` = True, if the</span>
-<span class="sd"> images are not evenly divisible by the ``tile_size``. The image data is</span>
-<span class="sd"> padded with a constant padding value of zero. For each image, the</span>
-<span class="sd"> corresponding ground truth image is padded with a "no data" label.</span>
-<span class="sd"> The default is False.</span>
-<span class="sd"> gt_pattern : `str`, optional</span>
-<span class="sd"> A regural expression to match the ground truth naming convention.</span>
-<span class="sd"> All directories and subdirectories in ``root_dir`` are searched for</span>
-<span class="sd"> files matching ``gt_pattern``. The default is '(.*)gt\\.tif'.</span>
-<span class="sd"> sort : `bool`, optional</span>
-<span class="sd"> Whether to chronologically sort the samples. Useful for time series</span>
-<span class="sd"> data. The default is False.</span>
-<span class="sd"> seed : `int`, optional</span>
-<span class="sd"> The random seed. Used to split the dataset into training, validation</span>
-<span class="sd"> and test set. Useful for reproducibility. The default is 0.</span>
-<span class="sd"> transforms : `list` [`pysegcnn.core.split.Augment`], optional</span>
-<span class="sd"> List of `pysegcnn.core.split.Augment` instances. Each item in</span>
-<span class="sd"> ``transforms`` generates a distinct transformed version of the dataset.</span>
-<span class="sd"> The total dataset is composed of the original untransformed dataset</span>
-<span class="sd"> together with each transformed version of it.</span>
-<span class="sd"> If ``transforms`` = [], only the original dataset is used.</span>
-<span class="sd"> The default is [].</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">root_dir</span><span class="p">,</span> <span class="n">use_bands</span><span class="o">=</span><span class="p">[],</span> <span class="n">tile_size</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">pad</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span>
- <span class="n">gt_pattern</span><span class="o">=</span><span class="s1">'(.*)gt.tif'</span><span class="p">,</span> <span class="n">sort</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">0</span><span class="p">,</span> <span class="n">transforms</span><span class="o">=</span><span class="p">[]):</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">()</span>
-
- <span class="c1"># dataset configuration</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">root</span> <span class="o">=</span> <span class="n">root_dir</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">use_bands</span> <span class="o">=</span> <span class="n">use_bands</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">tile_size</span> <span class="o">=</span> <span class="n">tile_size</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">pad</span> <span class="o">=</span> <span class="n">pad</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">gt_pattern</span> <span class="o">=</span> <span class="n">gt_pattern</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">sort</span> <span class="o">=</span> <span class="n">sort</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">seed</span> <span class="o">=</span> <span class="n">seed</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">transforms</span> <span class="o">=</span> <span class="n">transforms</span>
-
- <span class="c1"># initialize instance attributes</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">_init_attributes</span><span class="p">()</span>
-
- <span class="c1"># the samples of the dataset</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">scenes</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">compose_scenes</span><span class="p">()</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">_assert_compose_scenes</span><span class="p">()</span>
-
- <span class="k">def</span> <span class="nf">_init_attributes</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Initialize the class instance attributes."""</span>
- <span class="c1"># the size of a scene/patch in the dataset</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">size</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">get_size</span><span class="p">()</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">_assert_get_size</span><span class="p">()</span>
-
- <span class="c1"># the available spectral bands in the dataset</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">sensor</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">get_sensor</span><span class="p">()</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">_assert_get_sensor</span><span class="p">()</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">bands</span> <span class="o">=</span> <span class="p">{</span><span class="n">band</span><span class="o">.</span><span class="n">value</span><span class="p">:</span> <span class="n">band</span><span class="o">.</span><span class="n">name</span> <span class="k">for</span> <span class="n">band</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">sensor</span><span class="p">}</span>
-
- <span class="c1"># the class labels</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">_label_class</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">get_labels</span><span class="p">()</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">_assert_get_labels</span><span class="p">()</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">labels</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">_build_labels</span><span class="p">()</span>
-
- <span class="c1"># check which bands to use</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">use_bands</span> <span class="o">=</span> <span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">use_bands</span> <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">use_bands</span> <span class="k">else</span>
- <span class="p">[</span><span class="o">*</span><span class="bp">self</span><span class="o">.</span><span class="n">bands</span><span class="o">.</span><span class="n">values</span><span class="p">()])</span>
-
- <span class="c1"># calculate number of resulting tiles and check whether the images are</span>
- <span class="c1"># evenly divisible in square tiles of size (tile_size x tile_size)</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">tiles</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">padding</span> <span class="o">=</span> <span class="mi">1</span><span class="p">,</span> <span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">0</span><span class="p">)</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">tile_size</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">tiles</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">padding</span> <span class="o">=</span> <span class="n">is_divisible</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">size</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">tile_size</span><span class="p">,</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">pad</span><span class="p">)</span>
-
- <span class="c1"># the size of the padded scenes</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">height</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">size</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">+</span> <span class="bp">self</span><span class="o">.</span><span class="n">padding</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">+</span> <span class="bp">self</span><span class="o">.</span><span class="n">padding</span><span class="p">[</span><span class="mi">2</span><span class="p">]</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">width</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">size</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">+</span> <span class="bp">self</span><span class="o">.</span><span class="n">padding</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">+</span> <span class="bp">self</span><span class="o">.</span><span class="n">padding</span><span class="p">[</span><span class="mi">3</span><span class="p">]</span>
-
- <span class="c1"># the topleft corners of the tiles</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">tile_size</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">topleft</span> <span class="o">=</span> <span class="n">tile_topleft_corner</span><span class="p">((</span><span class="bp">self</span><span class="o">.</span><span class="n">height</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">width</span><span class="p">),</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">tile_size</span><span class="p">)</span>
-
- <span class="c1"># always use the original dataset together with the augmentations</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">transforms</span> <span class="o">=</span> <span class="p">[</span><span class="kc">None</span><span class="p">]</span> <span class="o">+</span> <span class="bp">self</span><span class="o">.</span><span class="n">transforms</span>
-
- <span class="c1"># when padding, add a new "no data" label to the ground truth</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">cval</span> <span class="o">=</span> <span class="mi">0</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">pad</span> <span class="ow">and</span> <span class="nb">sum</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">padding</span><span class="p">)</span> <span class="o">></span> <span class="mi">0</span><span class="p">:</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">cval</span> <span class="o">=</span> <span class="nb">max</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">labels</span><span class="p">)</span> <span class="o">+</span> <span class="mi">1</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">labels</span><span class="p">[</span><span class="bp">self</span><span class="o">.</span><span class="n">cval</span><span class="p">]</span> <span class="o">=</span> <span class="p">{</span><span class="s1">'label'</span><span class="p">:</span> <span class="s1">'No data'</span><span class="p">,</span> <span class="s1">'color'</span><span class="p">:</span> <span class="s1">'black'</span><span class="p">}</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Adding label "No data" with value=</span><span class="si">{}</span><span class="s1"> to ground truth.'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">cval</span><span class="p">))</span>
-
- <span class="k">def</span> <span class="nf">_build_labels</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Build the label dictionary.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> labels : `dict` [`int`, `dict`]</span>
-<span class="sd"> The label dictionary. The keys are the values of the class labels</span>
-<span class="sd"> in the ground truth ``y``. Each nested `dict` should have keys:</span>
-<span class="sd"> ``'color'``</span>
-<span class="sd"> A named color (`str`).</span>
-<span class="sd"> ``'label'``</span>
-<span class="sd"> The name of the class label (`str`).</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="p">{</span><span class="n">band</span><span class="o">.</span><span class="n">id</span><span class="p">:</span> <span class="p">{</span><span class="s1">'label'</span><span class="p">:</span> <span class="n">band</span><span class="o">.</span><span class="n">name</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="s1">'_'</span><span class="p">,</span> <span class="s1">' '</span><span class="p">),</span>
- <span class="s1">'color'</span><span class="p">:</span> <span class="n">band</span><span class="o">.</span><span class="n">color</span><span class="p">}</span>
- <span class="k">for</span> <span class="n">band</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">_label_class</span><span class="p">}</span>
-
- <span class="k">def</span> <span class="nf">_assert_compose_scenes</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Check whether compose_scenes() is correctly implemented."""</span>
- <span class="c1"># list of required keys</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">keys</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">use_bands</span> <span class="o">+</span> <span class="p">[</span><span class="s1">'gt'</span><span class="p">,</span> <span class="s1">'date'</span><span class="p">,</span> <span class="s1">'tile'</span><span class="p">,</span> <span class="s1">'transform'</span><span class="p">,</span> <span class="s1">'id'</span><span class="p">]</span>
-
- <span class="c1"># check if each scene is correctly composed</span>
- <span class="k">for</span> <span class="n">scene</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">scenes</span><span class="p">:</span>
- <span class="c1"># check the type of each scene</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">scene</span><span class="p">,</span> <span class="nb">dict</span><span class="p">):</span>
- <span class="k">raise</span> <span class="ne">TypeError</span><span class="p">(</span><span class="s1">'</span><span class="si">{}</span><span class="s1">.compose_scenes() should return a list of '</span>
- <span class="s1">'dictionaries.'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="vm">__class__</span><span class="o">.</span><span class="vm">__name__</span><span class="p">)</span>
- <span class="p">)</span>
-
- <span class="c1"># check if each scene dictionary has the correct keys</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="nb">all</span><span class="p">([</span><span class="n">k</span> <span class="ow">in</span> <span class="n">scene</span> <span class="k">for</span> <span class="n">k</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">keys</span><span class="p">]):</span>
- <span class="k">raise</span> <span class="ne">KeyError</span><span class="p">(</span><span class="s1">'Each scene dictionary should have keys </span><span class="si">{}</span><span class="s1">.'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">keys</span><span class="p">))</span>
-
- <span class="k">def</span> <span class="nf">_assert_get_size</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Check whether get_size() is correctly implemented."""</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="nb">isinstance</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">size</span><span class="p">,</span> <span class="nb">tuple</span><span class="p">)</span> <span class="ow">and</span> <span class="nb">len</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">size</span><span class="p">)</span> <span class="o">==</span> <span class="mi">2</span><span class="p">:</span>
- <span class="k">raise</span> <span class="ne">TypeError</span><span class="p">(</span><span class="s1">'</span><span class="si">{}</span><span class="s1">.get_size() should return the spatial size of '</span>
- <span class="s1">'an image sample as tuple, i.e. (height, width).'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="vm">__class__</span><span class="o">.</span><span class="vm">__name__</span><span class="p">))</span>
-
- <span class="k">def</span> <span class="nf">_assert_get_sensor</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Check whether get_sensor() is correctly implemented."""</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="nb">isinstance</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">sensor</span><span class="p">,</span> <span class="n">enum</span><span class="o">.</span><span class="n">EnumMeta</span><span class="p">):</span>
- <span class="k">raise</span> <span class="ne">TypeError</span><span class="p">(</span><span class="s1">'</span><span class="si">{}</span><span class="s1">.get_sensor() should return an instance of '</span>
- <span class="s1">'enum.Enum, containing an enumeration of the '</span>
- <span class="s1">'spectral bands of the sensor the dataset is '</span>
- <span class="s1">'derived from. Examples can be found in '</span>
- <span class="s1">'pysegcnn.core.constants.py.'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="vm">__class__</span><span class="o">.</span><span class="vm">__name__</span><span class="p">))</span>
-
- <span class="k">def</span> <span class="nf">_assert_get_labels</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Check whether get_labels() is correctly implemented."""</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="nb">issubclass</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">_label_class</span><span class="p">,</span> <span class="n">Label</span><span class="p">):</span>
- <span class="k">raise</span> <span class="ne">TypeError</span><span class="p">(</span><span class="s1">'</span><span class="si">{}</span><span class="s1">.get_labels() should return an instance of '</span>
- <span class="s1">'pysegcnn.core.constants.Label, '</span>
- <span class="s1">'containing an enumeration of the '</span>
- <span class="s1">'class labels, together with the corresponing id '</span>
- <span class="s1">'in the ground truth mask and a color for '</span>
- <span class="s1">'visualization. Examples can be found in '</span>
- <span class="s1">'pysegcnn.core.constants.py.'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="vm">__class__</span><span class="o">.</span><span class="vm">__name__</span><span class="p">))</span>
-
- <span class="k">def</span> <span class="fm">__len__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Return the number of samples in the dataset.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> nsamples : `int`</span>
-<span class="sd"> The number of samples in the dataset.</span>
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="nb">len</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">scenes</span><span class="p">)</span>
-
- <span class="k">def</span> <span class="fm">__getitem__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">idx</span><span class="p">):</span>
- <span class="sd">"""Return the data of a sample of the dataset given an index ``idx``.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> idx : `int`</span>
-<span class="sd"> The index of the sample.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> x : `torch.Tensor`</span>
-<span class="sd"> The sample input data.</span>
-<span class="sd"> y : `torch.Tensor`</span>
-<span class="sd"> The sample ground truth.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># select a scene</span>
- <span class="n">scene</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">read_scene</span><span class="p">(</span><span class="n">idx</span><span class="p">)</span>
-
- <span class="c1"># get samples</span>
- <span class="c1"># data: (tiles, bands, height, width)</span>
- <span class="c1"># gt: (height, width)</span>
- <span class="n">data</span><span class="p">,</span> <span class="n">gt</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">build_samples</span><span class="p">(</span><span class="n">scene</span><span class="p">)</span>
-
- <span class="c1"># preprocess samples</span>
- <span class="n">x</span><span class="p">,</span> <span class="n">y</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">preprocess</span><span class="p">(</span><span class="n">data</span><span class="p">,</span> <span class="n">gt</span><span class="p">)</span>
-
- <span class="c1"># apply transformation</span>
- <span class="k">if</span> <span class="n">scene</span><span class="p">[</span><span class="s1">'transform'</span><span class="p">]</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
- <span class="n">x</span><span class="p">,</span> <span class="n">y</span> <span class="o">=</span> <span class="n">scene</span><span class="p">[</span><span class="s1">'transform'</span><span class="p">](</span><span class="n">x</span><span class="p">,</span> <span class="n">y</span><span class="p">)</span>
-
- <span class="c1"># convert to torch tensors</span>
- <span class="n">x</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">to_tensor</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">torch</span><span class="o">.</span><span class="n">float32</span><span class="p">)</span>
- <span class="n">y</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">to_tensor</span><span class="p">(</span><span class="n">y</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">torch</span><span class="o">.</span><span class="n">uint8</span><span class="p">)</span>
-
- <span class="k">return</span> <span class="n">x</span><span class="p">,</span> <span class="n">y</span>
-
-<div class="viewcode-block" id="ImageDataset.compose_scenes"><a class="viewcode-back" href="../../source/core.html#core.dataset.ImageDataset.compose_scenes">[docs]</a> <span class="k">def</span> <span class="nf">compose_scenes</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Build the list of samples of the dataset.</span>
-
-<span class="sd"> Each sample is represented by a dictionary.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> NotImplementedError</span>
-<span class="sd"> Raised if the `pysegcnn.core.dataset.ImageDataset` class is not</span>
-<span class="sd"> inherited.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> samples : `list` [`dict`]</span>
-<span class="sd"> Each dictionary representing a sample should have keys:</span>
-<span class="sd"> ``'band_name_1'``</span>
-<span class="sd"> Path to the file of band_1.</span>
-<span class="sd"> ``'band_name_2'``</span>
-<span class="sd"> Path to the file of band_2.</span>
-<span class="sd"> ``'band_name_n'``</span>
-<span class="sd"> Path to the file of band_n.</span>
-<span class="sd"> ``'gt'``</span>
-<span class="sd"> Path to the ground truth file.</span>
-<span class="sd"> ``'date'``</span>
-<span class="sd"> The date of the sample.</span>
-<span class="sd"> ``'tile'``</span>
-<span class="sd"> The tile id of the sample.</span>
-<span class="sd"> ``'transform'``</span>
-<span class="sd"> The transformation to apply.</span>
-<span class="sd"> ``'id'``</span>
-<span class="sd"> The scene identifier.</span>
-
-<span class="sd"> """</span>
- <span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">(</span><span class="s1">'Inherit the ImageDataset class and '</span>
- <span class="s1">'implement the method.'</span><span class="p">)</span></div>
-
-<div class="viewcode-block" id="ImageDataset.get_size"><a class="viewcode-back" href="../../source/core.html#core.dataset.ImageDataset.get_size">[docs]</a> <span class="k">def</span> <span class="nf">get_size</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Return the size of the images in the dataset.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> NotImplementedError</span>
-<span class="sd"> Raised if the `pysegcnn.core.dataset.ImageDataset` class is not</span>
-<span class="sd"> inherited.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> size : `tuple`</span>
-<span class="sd"> The image size (height, width).</span>
-
-<span class="sd"> """</span>
- <span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">(</span><span class="s1">'Inherit the ImageDataset class and '</span>
- <span class="s1">'implement the method.'</span><span class="p">)</span></div>
-
-<div class="viewcode-block" id="ImageDataset.get_sensor"><a class="viewcode-back" href="../../source/core.html#core.dataset.ImageDataset.get_sensor">[docs]</a> <span class="k">def</span> <span class="nf">get_sensor</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Return an enumeration of the bands of the sensor of the dataset.</span>
-
-<span class="sd"> Examples can be found in `pysegcnn.core.constants`.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> NotImplementedError</span>
-<span class="sd"> Raised if the `pysegcnn.core.dataset.ImageDataset` class is not</span>
-<span class="sd"> inherited.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> sensor : `enum.Enum`</span>
-<span class="sd"> An enumeration of the bands of the sensor.</span>
-
-<span class="sd"> """</span>
- <span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">(</span><span class="s1">'Inherit the ImageDataset class and '</span>
- <span class="s1">'implement the method.'</span><span class="p">)</span></div>
-
-<div class="viewcode-block" id="ImageDataset.get_labels"><a class="viewcode-back" href="../../source/core.html#core.dataset.ImageDataset.get_labels">[docs]</a> <span class="k">def</span> <span class="nf">get_labels</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Return an enumeration of the class labels of the dataset.</span>
-
-<span class="sd"> Examples can be found in `pysegcnn.core.constants`.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> NotImplementedError</span>
-<span class="sd"> Raised if the `pysegcnn.core.dataset.ImageDataset` class is not</span>
-<span class="sd"> inherited.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> labels : `enum.Enum`</span>
-<span class="sd"> The class labels.</span>
-
-<span class="sd"> """</span>
- <span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">(</span><span class="s1">'Inherit the ImageDataset class and '</span>
- <span class="s1">'implement the method.'</span><span class="p">)</span></div>
-
-<div class="viewcode-block" id="ImageDataset.preprocess"><a class="viewcode-back" href="../../source/core.html#core.dataset.ImageDataset.preprocess">[docs]</a> <span class="k">def</span> <span class="nf">preprocess</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">data</span><span class="p">,</span> <span class="n">gt</span><span class="p">):</span>
- <span class="sd">"""Preprocess a sample before feeding it to a model.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> data : `numpy.ndarray`</span>
-<span class="sd"> The sample input data.</span>
-<span class="sd"> gt : `numpy.ndarray`</span>
-<span class="sd"> The sample ground truth.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> NotImplementedError</span>
-<span class="sd"> Raised if the `pysegcnn.core.dataset.ImageDataset` class is not</span>
-<span class="sd"> inherited.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> data : `numpy.ndarray`</span>
-<span class="sd"> The preprocessed input data.</span>
-<span class="sd"> gt : `numpy.ndarray`</span>
-<span class="sd"> The preprocessed ground truth data.</span>
-
-<span class="sd"> """</span>
- <span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">(</span><span class="s1">'Inherit the ImageDataset class and '</span>
- <span class="s1">'implement the method.'</span><span class="p">)</span></div>
-
-<div class="viewcode-block" id="ImageDataset.parse_scene_id"><a class="viewcode-back" href="../../source/core.html#core.dataset.ImageDataset.parse_scene_id">[docs]</a> <span class="k">def</span> <span class="nf">parse_scene_id</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">scene_id</span><span class="p">):</span>
- <span class="sd">"""Parse the scene identifier.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> scene_id : `str`</span>
-<span class="sd"> A scene identifier.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> NotImplementedError</span>
-<span class="sd"> Raised if the `pysegcnn.core.dataset.ImageDataset` class is not</span>
-<span class="sd"> inherited.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> scene : `dict` or `None`</span>
-<span class="sd"> A dictionary containing scene metadata. If `None`, ``scene_id`` is</span>
-<span class="sd"> not a valid scene identifier.</span>
-
-<span class="sd"> """</span>
- <span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">(</span><span class="s1">'Inherit the ImageDataset class and '</span>
- <span class="s1">'implement the method.'</span><span class="p">)</span></div>
-
-<div class="viewcode-block" id="ImageDataset.read_scene"><a class="viewcode-back" href="../../source/core.html#core.dataset.ImageDataset.read_scene">[docs]</a> <span class="k">def</span> <span class="nf">read_scene</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">idx</span><span class="p">):</span>
- <span class="sd">"""Read the data of the sample with index ``idx``.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> idx : `int`</span>
-<span class="sd"> The index of the sample.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> scene_data : `dict`</span>
-<span class="sd"> The sample data dictionary with keys:</span>
-<span class="sd"> ``'band_name_1'``</span>
-<span class="sd"> data of band_1 (`numpy.ndarray`).</span>
-<span class="sd"> ``'band_name_2'``</span>
-<span class="sd"> data of band_2 (`numpy.ndarray`).</span>
-<span class="sd"> ``'band_name_n'``</span>
-<span class="sd"> data of band_n (`numpy.ndarray`).</span>
-<span class="sd"> ``'gt'``</span>
-<span class="sd"> data of the ground truth (`numpy.ndarray`).</span>
-<span class="sd"> ``'date'``</span>
-<span class="sd"> The date of the sample.</span>
-<span class="sd"> ``'tile'``</span>
-<span class="sd"> The tile id of the sample.</span>
-<span class="sd"> ``'transform'``</span>
-<span class="sd"> The transformation to apply.</span>
-<span class="sd"> ``'id'``</span>
-<span class="sd"> The scene identifier.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># select a scene from the root directory</span>
- <span class="n">scene</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">scenes</span><span class="p">[</span><span class="n">idx</span><span class="p">]</span>
-
- <span class="c1"># read each band of the scene into a numpy array</span>
- <span class="n">scene_data</span> <span class="o">=</span> <span class="p">{}</span>
- <span class="k">for</span> <span class="n">key</span><span class="p">,</span> <span class="n">value</span> <span class="ow">in</span> <span class="n">scene</span><span class="o">.</span><span class="n">items</span><span class="p">():</span>
-
- <span class="c1"># pad zeros to each band if self.pad=True, but pad self.cval to</span>
- <span class="c1"># the ground truth</span>
- <span class="n">npad</span> <span class="o">=</span> <span class="mi">1</span> <span class="k">if</span> <span class="n">key</span> <span class="o">==</span> <span class="s1">'gt'</span> <span class="k">else</span> <span class="mi">0</span>
- <span class="k">if</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">value</span><span class="p">,</span> <span class="nb">str</span><span class="p">)</span> <span class="ow">and</span> <span class="n">key</span> <span class="o">!=</span> <span class="s1">'id'</span><span class="p">:</span>
- <span class="n">scene_data</span><span class="p">[</span><span class="n">key</span><span class="p">]</span> <span class="o">=</span> <span class="n">img2np</span><span class="p">(</span><span class="n">value</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">tile_size</span><span class="p">,</span> <span class="n">scene</span><span class="p">[</span><span class="s1">'tile'</span><span class="p">],</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">pad</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">cval</span> <span class="o">*</span> <span class="n">npad</span><span class="p">)</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="n">scene_data</span><span class="p">[</span><span class="n">key</span><span class="p">]</span> <span class="o">=</span> <span class="n">value</span>
-
- <span class="k">return</span> <span class="n">scene_data</span></div>
-
-<div class="viewcode-block" id="ImageDataset.build_samples"><a class="viewcode-back" href="../../source/core.html#core.dataset.ImageDataset.build_samples">[docs]</a> <span class="k">def</span> <span class="nf">build_samples</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">scene</span><span class="p">):</span>
- <span class="sd">"""Stack the bands of a sample in a single array.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> scene : `dict`</span>
-<span class="sd"> The sample data dictionary with keys:</span>
-<span class="sd"> ``'band_name_1'``</span>
-<span class="sd"> data of band_1 (`numpy.ndarray`).</span>
-<span class="sd"> ``'band_name_2'``</span>
-<span class="sd"> data of band_2 (`numpy.ndarray`).</span>
-<span class="sd"> ``'band_name_n'``</span>
-<span class="sd"> data of band_n (`numpy.ndarray`).</span>
-<span class="sd"> ``'gt'``</span>
-<span class="sd"> data of the ground truth (`numpy.ndarray`).</span>
-<span class="sd"> ``'date'``</span>
-<span class="sd"> The date of the sample.</span>
-<span class="sd"> ``'tile'``</span>
-<span class="sd"> The tile id of the sample.</span>
-<span class="sd"> ``'transform'``</span>
-<span class="sd"> The transformation to apply.</span>
-<span class="sd"> ``'id'``</span>
-<span class="sd"> The scene identifier.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> stack : `numpy.ndarray`</span>
-<span class="sd"> The input data of the sample.</span>
-<span class="sd"> gt : TYPE</span>
-<span class="sd"> The ground truth of the sample.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># iterate over the channels to stack</span>
- <span class="n">stack</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">stack</span><span class="p">([</span><span class="n">scene</span><span class="p">[</span><span class="n">band</span><span class="p">]</span> <span class="k">for</span> <span class="n">band</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">use_bands</span><span class="p">],</span> <span class="n">axis</span><span class="o">=</span><span class="mi">0</span><span class="p">)</span>
- <span class="n">gt</span> <span class="o">=</span> <span class="n">scene</span><span class="p">[</span><span class="s1">'gt'</span><span class="p">]</span>
-
- <span class="k">return</span> <span class="n">stack</span><span class="p">,</span> <span class="n">gt</span></div>
-
-<div class="viewcode-block" id="ImageDataset.to_tensor"><a class="viewcode-back" href="../../source/core.html#core.dataset.ImageDataset.to_tensor">[docs]</a> <span class="k">def</span> <span class="nf">to_tensor</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">x</span><span class="p">,</span> <span class="n">dtype</span><span class="p">):</span>
- <span class="sd">"""Convert ``x`` to `torch.Tensor`.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> x : array_like</span>
-<span class="sd"> The input data.</span>
-<span class="sd"> dtype : `torch.dtype`</span>
-<span class="sd"> The data type used to convert ``x``.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> x : `torch.Tensor`</span>
-<span class="sd"> The input data tensor.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="n">torch</span><span class="o">.</span><span class="n">tensor</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">x</span><span class="p">)</span><span class="o">.</span><span class="n">copy</span><span class="p">(),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">dtype</span><span class="p">)</span></div>
-
- <span class="k">def</span> <span class="fm">__repr__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Dataset representation.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> fs : `str`</span>
-<span class="sd"> Representation string.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># representation string to print</span>
- <span class="n">fs</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="vm">__class__</span><span class="o">.</span><span class="vm">__name__</span> <span class="o">+</span> <span class="s1">'(</span><span class="se">\n</span><span class="s1">'</span>
-
- <span class="c1"># sensor</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">' (sensor):</span><span class="se">\n</span><span class="s1"> - '</span> <span class="o">+</span> <span class="bp">self</span><span class="o">.</span><span class="n">sensor</span><span class="o">.</span><span class="vm">__name__</span>
-
- <span class="c1"># bands used for the segmentation</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">'</span><span class="se">\n</span><span class="s1"> (bands):</span><span class="se">\n</span><span class="s1"> '</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">'</span><span class="se">\n</span><span class="s1"> '</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="s1">'- Band </span><span class="si">{}</span><span class="s1">: </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">i</span><span class="p">,</span> <span class="n">b</span><span class="p">)</span> <span class="k">for</span> <span class="n">i</span><span class="p">,</span> <span class="n">b</span> <span class="ow">in</span>
- <span class="nb">enumerate</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">use_bands</span><span class="p">))</span>
-
- <span class="c1"># scenes</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">'</span><span class="se">\n</span><span class="s1"> (scene):</span><span class="se">\n</span><span class="s1"> '</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">'- size (h, w): </span><span class="si">{}</span><span class="se">\n</span><span class="s1"> '</span><span class="o">.</span><span class="n">format</span><span class="p">((</span><span class="bp">self</span><span class="o">.</span><span class="n">height</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">width</span><span class="p">))</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">'- number of scenes: </span><span class="si">{}</span><span class="se">\n</span><span class="s1"> '</span><span class="o">.</span><span class="n">format</span><span class="p">(</span>
- <span class="nb">len</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">unique</span><span class="p">([</span><span class="n">f</span><span class="p">[</span><span class="s1">'id'</span><span class="p">]</span> <span class="k">for</span> <span class="n">f</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">scenes</span><span class="p">])))</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">'- padding (bottom, left, top, right): </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">padding</span><span class="p">)</span>
-
- <span class="c1"># tiles</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">'</span><span class="se">\n</span><span class="s1"> (tiles):</span><span class="se">\n</span><span class="s1"> '</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">'- number of tiles per scene: </span><span class="si">{}</span><span class="se">\n</span><span class="s1"> '</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">tiles</span><span class="p">)</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">'- tile size: </span><span class="si">{}</span><span class="se">\n</span><span class="s1"> '</span><span class="o">.</span><span class="n">format</span><span class="p">((</span><span class="bp">self</span><span class="o">.</span><span class="n">tile_size</span><span class="p">,</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">tile_size</span><span class="p">))</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">'- number of tiles: </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="nb">len</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">scenes</span><span class="p">))</span>
-
- <span class="c1"># classes of interest</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">'</span><span class="se">\n</span><span class="s1"> (classes):</span><span class="se">\n</span><span class="s1"> '</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">'</span><span class="se">\n</span><span class="s1"> '</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="s1">'- Class </span><span class="si">{}</span><span class="s1">: </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">k</span><span class="p">,</span> <span class="n">v</span><span class="p">[</span><span class="s1">'label'</span><span class="p">])</span> <span class="k">for</span>
- <span class="n">k</span><span class="p">,</span> <span class="n">v</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">labels</span><span class="o">.</span><span class="n">items</span><span class="p">())</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">'</span><span class="se">\n</span><span class="s1">)'</span>
- <span class="k">return</span> <span class="n">fs</span></div>
-
-
-<div class="viewcode-block" id="StandardEoDataset"><a class="viewcode-back" href="../../source/core.html#core.dataset.StandardEoDataset">[docs]</a><span class="k">class</span> <span class="nc">StandardEoDataset</span><span class="p">(</span><span class="n">ImageDataset</span><span class="p">):</span>
- <span class="sa">r</span><span class="sd">"""Base class for standard Earth Observation style datasets.</span>
-
-<span class="sd"> `pysegcnn.core.dataset.StandardEoDataset` implements the</span>
-<span class="sd"> `~pysegcnn.core.dataset.StandardEoDataset.compose_scenes` method for</span>
-<span class="sd"> datasets with the following directory structure:</span>
-
-<span class="sd"> root_dir/</span>
-<span class="sd"> scene_id_1/</span>
-<span class="sd"> scene_id_1_B1.tif</span>
-<span class="sd"> scene_id_1_B2.tif</span>
-<span class="sd"> .</span>
-<span class="sd"> .</span>
-<span class="sd"> .</span>
-<span class="sd"> scene_id_1_BN.tif</span>
-<span class="sd"> scene_id_2/</span>
-<span class="sd"> scene_id_2_B1.tif</span>
-<span class="sd"> scene_id_2_B2.tif</span>
-<span class="sd"> .</span>
-<span class="sd"> .</span>
-<span class="sd"> .</span>
-<span class="sd"> scene_id_2_BN.tif</span>
-<span class="sd"> .</span>
-<span class="sd"> .</span>
-<span class="sd"> .</span>
-<span class="sd"> scene_id_N/</span>
-<span class="sd"> .</span>
-<span class="sd"> .</span>
-<span class="sd"> .</span>
-
-<span class="sd"> If your dataset shares this directory structure, you can directly inherit</span>
-<span class="sd"> `pysegcnn.core.dataset.StandardEoDataset` and implement the remaining</span>
-<span class="sd"> methods.</span>
-
-<span class="sd"> See `pysegcnn.core.dataset.SparcsDataset` for an example.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> root_dir : `str`</span>
-<span class="sd"> The root directory, path to the dataset.</span>
-<span class="sd"> use_bands : `list` [`str`], optional</span>
-<span class="sd"> A list of the spectral bands to use. The default is [].</span>
-<span class="sd"> tile_size : `int` or `None`, optional</span>
-<span class="sd"> The size of the tiles. If not `None`, each scene is divided into square</span>
-<span class="sd"> tiles of shape (tile_size, tile_size). The default is None.</span>
-<span class="sd"> pad : `bool`, optional</span>
-<span class="sd"> Whether to center pad the input image. Set ``pad`` = True, if the</span>
-<span class="sd"> images are not evenly divisible by the ``tile_size``. The image data is</span>
-<span class="sd"> padded with a constant padding value of zero. For each image, the</span>
-<span class="sd"> corresponding ground truth image is padded with a "no data" label.</span>
-<span class="sd"> The default is False.</span>
-<span class="sd"> gt_pattern : `str`, optional</span>
-<span class="sd"> A regural expression to match the ground truth naming convention. All</span>
-<span class="sd"> directories and subdirectories in ``root_dir`` are searched for files</span>
-<span class="sd"> matching ``gt_pattern``. The default is '(.*)gt\\.tif'.</span>
-<span class="sd"> sort : `bool`, optional</span>
-<span class="sd"> Whether to chronologically sort the samples. Useful for time series</span>
-<span class="sd"> data. The default is False.</span>
-<span class="sd"> seed : `int`, optional</span>
-<span class="sd"> The random seed. Used to split the dataset into training, validation</span>
-<span class="sd"> and test set. Useful for reproducibility. The default is 0.</span>
-<span class="sd"> transforms : `list` [`pysegcnn.core.split.Augment`], optional</span>
-<span class="sd"> List of `pysegcnn.core.split.Augment` instances. Each item in</span>
-<span class="sd"> ``transforms`` generates a distinct transformed version of the dataset.</span>
-<span class="sd"> The total dataset is composed of the original untransformed dataset</span>
-<span class="sd"> together with each transformed version of it.</span>
-<span class="sd"> If ``transforms`` = [], only the original dataset is used.</span>
-<span class="sd"> The default is [].</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">root_dir</span><span class="p">,</span> <span class="n">use_bands</span><span class="o">=</span><span class="p">[],</span> <span class="n">tile_size</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">pad</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span>
- <span class="n">gt_pattern</span><span class="o">=</span><span class="s1">'(.*)gt</span><span class="se">\\</span><span class="s1">.tif'</span><span class="p">,</span> <span class="n">sort</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">0</span><span class="p">,</span> <span class="n">transforms</span><span class="o">=</span><span class="p">[]):</span>
- <span class="c1"># initialize super class ImageDataset</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">root_dir</span><span class="p">,</span> <span class="n">use_bands</span><span class="p">,</span> <span class="n">tile_size</span><span class="p">,</span> <span class="n">pad</span><span class="p">,</span> <span class="n">gt_pattern</span><span class="p">,</span>
- <span class="n">sort</span><span class="p">,</span> <span class="n">seed</span><span class="p">,</span> <span class="n">transforms</span><span class="p">)</span>
-
- <span class="k">def</span> <span class="nf">_get_band_number</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">path</span><span class="p">):</span>
- <span class="sd">"""Return the band number of a scene .tif file.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> path : `str`</span>
-<span class="sd"> The path to the .tif file.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> band : `int` or `str`</span>
-<span class="sd"> The band number.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># filename</span>
- <span class="n">fname</span> <span class="o">=</span> <span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">basename</span><span class="p">(</span><span class="n">path</span><span class="p">)</span>
-
- <span class="c1"># search for numbers following a "B" in the filename</span>
- <span class="n">band</span> <span class="o">=</span> <span class="n">re</span><span class="o">.</span><span class="n">search</span><span class="p">(</span><span class="s1">'B</span><span class="se">\\</span><span class="s1">dA|B</span><span class="se">\\</span><span class="s1">d{1,2}'</span><span class="p">,</span> <span class="n">fname</span><span class="p">)[</span><span class="mi">0</span><span class="p">]</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="s1">'B'</span><span class="p">,</span> <span class="s1">''</span><span class="p">)</span>
-
- <span class="c1"># try converting to an integer:</span>
- <span class="c1"># raises a ValueError for Sentinel2 8A band</span>
- <span class="k">try</span><span class="p">:</span>
- <span class="n">band</span> <span class="o">=</span> <span class="nb">int</span><span class="p">(</span><span class="n">band</span><span class="p">)</span>
- <span class="k">except</span> <span class="ne">ValueError</span><span class="p">:</span>
- <span class="k">pass</span>
-
- <span class="k">return</span> <span class="n">band</span>
-
- <span class="k">def</span> <span class="nf">_store_bands</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">bands</span><span class="p">,</span> <span class="n">gt</span><span class="p">):</span>
- <span class="sd">"""Write the bands of interest to a dictionary.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> bands : `list` [`str`]</span>
-<span class="sd"> Paths to the .tif files of the bands of the scene.</span>
-<span class="sd"> gt : `str`</span>
-<span class="sd"> Path to the ground truth of the scene.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> scene_data : `dict`</span>
-<span class="sd"> The scene data dictionary with keys:</span>
-<span class="sd"> ``'band_name_1'``</span>
-<span class="sd"> Path to the .tif file of band_1.</span>
-<span class="sd"> ``'band_name_2'``</span>
-<span class="sd"> Path to the .tif file of band_2.</span>
-<span class="sd"> ``'band_name_n'``</span>
-<span class="sd"> Path to the .tif file of band_n.</span>
-<span class="sd"> ``'gt'``</span>
-<span class="sd"> Path to the ground truth file.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># store the bands of interest in a dictionary</span>
- <span class="n">scene_data</span> <span class="o">=</span> <span class="p">{}</span>
- <span class="k">for</span> <span class="n">i</span><span class="p">,</span> <span class="n">b</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">(</span><span class="n">bands</span><span class="p">):</span>
- <span class="n">band</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">bands</span><span class="p">[</span><span class="bp">self</span><span class="o">.</span><span class="n">_get_band_number</span><span class="p">(</span><span class="n">b</span><span class="p">)]</span>
- <span class="k">if</span> <span class="n">band</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">use_bands</span><span class="p">:</span>
- <span class="n">scene_data</span><span class="p">[</span><span class="n">band</span><span class="p">]</span> <span class="o">=</span> <span class="n">b</span>
-
- <span class="c1"># store ground truth</span>
- <span class="n">scene_data</span><span class="p">[</span><span class="s1">'gt'</span><span class="p">]</span> <span class="o">=</span> <span class="n">gt</span>
-
- <span class="k">return</span> <span class="n">scene_data</span>
-
-<div class="viewcode-block" id="StandardEoDataset.compose_scenes"><a class="viewcode-back" href="../../source/core.html#core.dataset.StandardEoDataset.compose_scenes">[docs]</a> <span class="k">def</span> <span class="nf">compose_scenes</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Build the list of samples of the dataset.</span>
-
-<span class="sd"> Each sample is represented by a dictionary.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> scenes : `list` [`dict`]</span>
-<span class="sd"> Each item in ``scenes`` is a `dict` with keys:</span>
-<span class="sd"> ``'band_name_1'``</span>
-<span class="sd"> Path to the file of band_1.</span>
-<span class="sd"> ``'band_name_2'``</span>
-<span class="sd"> Path to the file of band_2.</span>
-<span class="sd"> ``'band_name_n'``</span>
-<span class="sd"> Path to the file of band_n.</span>
-<span class="sd"> ``'gt'``</span>
-<span class="sd"> Path to the ground truth file.</span>
-<span class="sd"> ``'date'``</span>
-<span class="sd"> The date of the sample.</span>
-<span class="sd"> ``'tile'``</span>
-<span class="sd"> The tile id of the sample.</span>
-<span class="sd"> ``'transform'``</span>
-<span class="sd"> The transformation to apply.</span>
-<span class="sd"> ``'id'``</span>
-<span class="sd"> The scene identifier.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># search the root directory</span>
- <span class="n">scenes</span> <span class="o">=</span> <span class="p">[]</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">gt</span> <span class="o">=</span> <span class="p">[]</span>
- <span class="k">for</span> <span class="n">dirpath</span><span class="p">,</span> <span class="n">dirname</span><span class="p">,</span> <span class="n">files</span> <span class="ow">in</span> <span class="n">os</span><span class="o">.</span><span class="n">walk</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">root</span><span class="p">):</span>
-
- <span class="c1"># search for a ground truth in the current directory</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">gt</span><span class="o">.</span><span class="n">extend</span><span class="p">([</span><span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">dirpath</span><span class="p">,</span> <span class="n">f</span><span class="p">)</span> <span class="k">for</span> <span class="n">f</span> <span class="ow">in</span> <span class="n">files</span>
- <span class="k">if</span> <span class="n">re</span><span class="o">.</span><span class="n">search</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">gt_pattern</span><span class="p">,</span> <span class="n">f</span><span class="p">)])</span>
-
- <span class="c1"># check if the current directory name matches a scene identifier</span>
- <span class="n">scene</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">parse_scene_id</span><span class="p">(</span><span class="n">dirpath</span><span class="p">)</span>
-
- <span class="k">if</span> <span class="n">scene</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
-
- <span class="c1"># get the date of the current scene</span>
- <span class="n">date</span> <span class="o">=</span> <span class="n">scene</span><span class="p">[</span><span class="s1">'date'</span><span class="p">]</span>
-
- <span class="c1"># list the spectral bands of the scene</span>
- <span class="n">bands</span> <span class="o">=</span> <span class="p">[</span><span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">dirpath</span><span class="p">,</span> <span class="n">f</span><span class="p">)</span> <span class="k">for</span> <span class="n">f</span> <span class="ow">in</span> <span class="n">files</span>
- <span class="k">if</span> <span class="n">re</span><span class="o">.</span><span class="n">search</span><span class="p">(</span><span class="s1">'B</span><span class="se">\\</span><span class="s1">dA|B</span><span class="se">\\</span><span class="s1">d{1,2}'</span><span class="p">,</span> <span class="n">f</span><span class="p">)]</span>
-
- <span class="c1"># get the ground truth mask</span>
- <span class="k">try</span><span class="p">:</span>
- <span class="n">gt</span> <span class="o">=</span> <span class="p">[</span><span class="n">truth</span> <span class="k">for</span> <span class="n">truth</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">gt</span> <span class="k">if</span> <span class="n">scene</span><span class="p">[</span><span class="s1">'id'</span><span class="p">]</span> <span class="ow">in</span> <span class="n">truth</span><span class="p">]</span>
- <span class="n">gt</span> <span class="o">=</span> <span class="n">gt</span><span class="o">.</span><span class="n">pop</span><span class="p">()</span>
-
- <span class="k">except</span> <span class="ne">IndexError</span><span class="p">:</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Skipping scene </span><span class="si">{}</span><span class="s1">: ground truth not available'</span>
- <span class="s1">' (pattern = </span><span class="si">{}</span><span class="s1">).'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">scene</span><span class="p">[</span><span class="s1">'id'</span><span class="p">],</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">gt_pattern</span><span class="p">))</span>
- <span class="k">continue</span>
-
- <span class="c1"># iterate over the tiles</span>
- <span class="k">for</span> <span class="n">tile</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">tiles</span><span class="p">):</span>
-
- <span class="c1"># iterate over the transformations to apply</span>
- <span class="k">for</span> <span class="n">transf</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">transforms</span><span class="p">:</span>
-
- <span class="c1"># store the bands and the ground truth mask of the tile</span>
- <span class="n">data</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">_store_bands</span><span class="p">(</span><span class="n">bands</span><span class="p">,</span> <span class="n">gt</span><span class="p">)</span>
-
- <span class="c1"># the name of the scene</span>
- <span class="n">data</span><span class="p">[</span><span class="s1">'id'</span><span class="p">]</span> <span class="o">=</span> <span class="n">scene</span><span class="p">[</span><span class="s1">'id'</span><span class="p">]</span>
-
- <span class="c1"># store tile number</span>
- <span class="n">data</span><span class="p">[</span><span class="s1">'tile'</span><span class="p">]</span> <span class="o">=</span> <span class="n">tile</span>
-
- <span class="c1"># store date</span>
- <span class="n">data</span><span class="p">[</span><span class="s1">'date'</span><span class="p">]</span> <span class="o">=</span> <span class="n">date</span>
-
- <span class="c1"># store optional transformation</span>
- <span class="n">data</span><span class="p">[</span><span class="s1">'transform'</span><span class="p">]</span> <span class="o">=</span> <span class="n">transf</span>
-
- <span class="c1"># append to list</span>
- <span class="n">scenes</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">data</span><span class="p">)</span>
-
- <span class="c1"># sort list of scenes and ground truths in chronological order</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">sort</span><span class="p">:</span>
- <span class="n">scenes</span><span class="o">.</span><span class="n">sort</span><span class="p">(</span><span class="n">key</span><span class="o">=</span><span class="k">lambda</span> <span class="n">k</span><span class="p">:</span> <span class="n">k</span><span class="p">[</span><span class="s1">'date'</span><span class="p">])</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">gt</span><span class="o">.</span><span class="n">sort</span><span class="p">(</span><span class="n">key</span><span class="o">=</span><span class="k">lambda</span> <span class="n">k</span><span class="p">:</span> <span class="bp">self</span><span class="o">.</span><span class="n">parse_scene_id</span><span class="p">(</span><span class="n">k</span><span class="p">)[</span><span class="s1">'date'</span><span class="p">])</span>
-
- <span class="k">return</span> <span class="n">scenes</span></div></div>
-
-
-<div class="viewcode-block" id="SparcsDataset"><a class="viewcode-back" href="../../source/core.html#core.dataset.SparcsDataset">[docs]</a><span class="k">class</span> <span class="nc">SparcsDataset</span><span class="p">(</span><span class="n">StandardEoDataset</span><span class="p">):</span>
- <span class="sa">r</span><span class="sd">"""Class for the `Sparcs`_ dataset.</span>
-
-<span class="sd"> .. _Sparcs:</span>
-<span class="sd"> https://www.usgs.gov/land-resources/nli/landsat/spatial-procedures-automated-removal-cloud-and-shadow-sparcs-validation</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> root_dir : `str`</span>
-<span class="sd"> The root directory, path to the dataset.</span>
-<span class="sd"> use_bands : `list` [`str`], optional</span>
-<span class="sd"> A list of the spectral bands to use. The default is [].</span>
-<span class="sd"> tile_size : `int` or `None`, optional</span>
-<span class="sd"> The size of the tiles. If not `None`, each scene is divided into square</span>
-<span class="sd"> tiles of shape (tile_size, tile_size). The default is None.</span>
-<span class="sd"> pad : `bool`, optional</span>
-<span class="sd"> Whether to center pad the input image. Set ``pad`` = True, if the</span>
-<span class="sd"> images are not evenly divisible by the ``tile_size``. The image data is</span>
-<span class="sd"> padded with a constant padding value of zero. For each image, the</span>
-<span class="sd"> corresponding ground truth image is padded with a "no data" label.</span>
-<span class="sd"> The default is False.</span>
-<span class="sd"> gt_pattern : `str`, optional</span>
-<span class="sd"> A regural expression to match the ground truth naming convention. All</span>
-<span class="sd"> directories and subdirectories in ``root_dir`` are searched for files</span>
-<span class="sd"> matching ``gt_pattern``. The default is '(.*)gt\\.tif'.</span>
-<span class="sd"> sort : `bool`, optional</span>
-<span class="sd"> Whether to chronologically sort the samples. Useful for time series</span>
-<span class="sd"> data. The default is False.</span>
-<span class="sd"> seed : `int`, optional</span>
-<span class="sd"> The random seed. Used to split the dataset into training, validation</span>
-<span class="sd"> and test set. Useful for reproducibility. The default is 0.</span>
-<span class="sd"> transforms : `list` [`pysegcnn.core.split.Augment`], optional</span>
-<span class="sd"> List of `pysegcnn.core.split.Augment` instances. Each item in</span>
-<span class="sd"> ``transforms`` generates a distinct transformed version of the dataset.</span>
-<span class="sd"> The total dataset is composed of the original untransformed dataset</span>
-<span class="sd"> together with each transformed version of it.</span>
-<span class="sd"> If ``transforms`` = [], only the original dataset is used.</span>
-<span class="sd"> The default is [].</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">root_dir</span><span class="p">,</span> <span class="n">use_bands</span><span class="o">=</span><span class="p">[],</span> <span class="n">tile_size</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">pad</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span>
- <span class="n">gt_pattern</span><span class="o">=</span><span class="s1">'(.*)gt</span><span class="se">\\</span><span class="s1">.tif'</span><span class="p">,</span> <span class="n">sort</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">0</span><span class="p">,</span> <span class="n">transforms</span><span class="o">=</span><span class="p">[]):</span>
- <span class="c1"># initialize super class StandardEoDataset</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">root_dir</span><span class="p">,</span> <span class="n">use_bands</span><span class="p">,</span> <span class="n">tile_size</span><span class="p">,</span> <span class="n">pad</span><span class="p">,</span> <span class="n">gt_pattern</span><span class="p">,</span>
- <span class="n">sort</span><span class="p">,</span> <span class="n">seed</span><span class="p">,</span> <span class="n">transforms</span><span class="p">)</span>
-
-<div class="viewcode-block" id="SparcsDataset.get_size"><a class="viewcode-back" href="../../source/core.html#core.dataset.SparcsDataset.get_size">[docs]</a> <span class="k">def</span> <span class="nf">get_size</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Image size of the Sparcs dataset.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> size : `tuple`</span>
-<span class="sd"> The image size (height, width).</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="p">(</span><span class="mi">1000</span><span class="p">,</span> <span class="mi">1000</span><span class="p">)</span></div>
-
-<div class="viewcode-block" id="SparcsDataset.get_sensor"><a class="viewcode-back" href="../../source/core.html#core.dataset.SparcsDataset.get_sensor">[docs]</a> <span class="k">def</span> <span class="nf">get_sensor</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Landsat 8 bands of the Sparcs dataset.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> sensor : `enum.Enum`</span>
-<span class="sd"> An enumeration of the bands of the sensor.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="n">Landsat8</span></div>
-
-<div class="viewcode-block" id="SparcsDataset.get_labels"><a class="viewcode-back" href="../../source/core.html#core.dataset.SparcsDataset.get_labels">[docs]</a> <span class="k">def</span> <span class="nf">get_labels</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Class labels of the Sparcs dataset.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> labels : `enum.Enum`</span>
-<span class="sd"> The class labels.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="n">SparcsLabels</span></div>
-
-<div class="viewcode-block" id="SparcsDataset.preprocess"><a class="viewcode-back" href="../../source/core.html#core.dataset.SparcsDataset.preprocess">[docs]</a> <span class="k">def</span> <span class="nf">preprocess</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">data</span><span class="p">,</span> <span class="n">gt</span><span class="p">):</span>
- <span class="sd">"""Preprocess Sparcs dataset images.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> data : `numpy.ndarray`</span>
-<span class="sd"> The sample input data.</span>
-<span class="sd"> gt : `numpy.ndarray`</span>
-<span class="sd"> The sample ground truth.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> data : `numpy.ndarray`</span>
-<span class="sd"> The preprocessed input data.</span>
-<span class="sd"> gt : `numpy.ndarray`</span>
-<span class="sd"> The preprocessed ground truth data.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># if the preprocessing is not done externally, implement it here</span>
- <span class="k">return</span> <span class="n">data</span><span class="p">,</span> <span class="n">gt</span></div>
-
-<div class="viewcode-block" id="SparcsDataset.parse_scene_id"><a class="viewcode-back" href="../../source/core.html#core.dataset.SparcsDataset.parse_scene_id">[docs]</a> <span class="k">def</span> <span class="nf">parse_scene_id</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">scene_id</span><span class="p">):</span>
- <span class="sd">"""Parse Sparcs scene identifiers (Landsat 8).</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> scene_id : `str`</span>
-<span class="sd"> A scene identifier.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> scene : `dict` or `None`</span>
-<span class="sd"> A dictionary containing scene metadata. If `None`, ``scene_id`` is</span>
-<span class="sd"> not a valid Landsat scene identifier.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="n">parse_landsat_scene</span><span class="p">(</span><span class="n">scene_id</span><span class="p">)</span></div></div>
-
-
-<div class="viewcode-block" id="ProSnowDataset"><a class="viewcode-back" href="../../source/core.html#core.dataset.ProSnowDataset">[docs]</a><span class="k">class</span> <span class="nc">ProSnowDataset</span><span class="p">(</span><span class="n">StandardEoDataset</span><span class="p">):</span>
- <span class="sa">r</span><span class="sd">"""Class for the ProSnow datasets.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> root_dir : `str`</span>
-<span class="sd"> The root directory, path to the dataset.</span>
-<span class="sd"> use_bands : `list` [`str`], optional</span>
-<span class="sd"> A list of the spectral bands to use. The default is [].</span>
-<span class="sd"> tile_size : `int` or `None`, optional</span>
-<span class="sd"> The size of the tiles. If not `None`, each scene is divided into square</span>
-<span class="sd"> tiles of shape (tile_size, tile_size). The default is None.</span>
-<span class="sd"> pad : `bool`, optional</span>
-<span class="sd"> Whether to center pad the input image. Set ``pad`` = True, if the</span>
-<span class="sd"> images are not evenly divisible by the ``tile_size``. The image data is</span>
-<span class="sd"> padded with a constant padding value of zero. For each image, the</span>
-<span class="sd"> corresponding ground truth image is padded with a "no data" label.</span>
-<span class="sd"> The default is False.</span>
-<span class="sd"> gt_pattern : `str`, optional</span>
-<span class="sd"> A regural expression to match the ground truth naming convention. All</span>
-<span class="sd"> directories and subdirectories in ``root_dir`` are searched for files</span>
-<span class="sd"> matching ``gt_pattern``. The default is '(.*)gt\\.tif'.</span>
-<span class="sd"> sort : `bool`, optional</span>
-<span class="sd"> Whether to chronologically sort the samples. Useful for time series</span>
-<span class="sd"> data. The default is False.</span>
-<span class="sd"> seed : `int`, optional</span>
-<span class="sd"> The random seed. Used to split the dataset into training, validation</span>
-<span class="sd"> and test set. Useful for reproducibility. The default is 0.</span>
-<span class="sd"> transforms : `list` [`pysegcnn.core.split.Augment`], optional</span>
-<span class="sd"> List of `pysegcnn.core.split.Augment` instances. Each item in</span>
-<span class="sd"> ``transforms`` generates a distinct transformed version of the dataset.</span>
-<span class="sd"> The total dataset is composed of the original untransformed dataset</span>
-<span class="sd"> together with each transformed version of it.</span>
-<span class="sd"> If ``transforms`` = [], only the original dataset is used.</span>
-<span class="sd"> The default is [].</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">root_dir</span><span class="p">,</span> <span class="n">use_bands</span><span class="o">=</span><span class="p">[],</span> <span class="n">tile_size</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">pad</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span>
- <span class="n">gt_pattern</span><span class="o">=</span><span class="s1">'(.*)gt</span><span class="se">\\</span><span class="s1">.tif'</span><span class="p">,</span> <span class="n">sort</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">0</span><span class="p">,</span> <span class="n">transforms</span><span class="o">=</span><span class="p">[]):</span>
- <span class="c1"># initialize super class StandardEoDataset</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">root_dir</span><span class="p">,</span> <span class="n">use_bands</span><span class="p">,</span> <span class="n">tile_size</span><span class="p">,</span> <span class="n">pad</span><span class="p">,</span> <span class="n">gt_pattern</span><span class="p">,</span>
- <span class="n">sort</span><span class="p">,</span> <span class="n">seed</span><span class="p">,</span> <span class="n">transforms</span><span class="p">)</span>
-
-<div class="viewcode-block" id="ProSnowDataset.get_sensor"><a class="viewcode-back" href="../../source/core.html#core.dataset.ProSnowDataset.get_sensor">[docs]</a> <span class="k">def</span> <span class="nf">get_sensor</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Sentinel 2 bands of the ProSnow datasets.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> sensor : `enum.Enum`</span>
-<span class="sd"> An enumeration of the bands of the sensor.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="n">Sentinel2</span></div>
-
-<div class="viewcode-block" id="ProSnowDataset.get_labels"><a class="viewcode-back" href="../../source/core.html#core.dataset.ProSnowDataset.get_labels">[docs]</a> <span class="k">def</span> <span class="nf">get_labels</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Class labels of the ProSnow datasets.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> labels : `enum.Enum`</span>
-<span class="sd"> The class labels.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="n">ProSnowLabels</span></div>
-
-<div class="viewcode-block" id="ProSnowDataset.preprocess"><a class="viewcode-back" href="../../source/core.html#core.dataset.ProSnowDataset.preprocess">[docs]</a> <span class="k">def</span> <span class="nf">preprocess</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">data</span><span class="p">,</span> <span class="n">gt</span><span class="p">):</span>
- <span class="sd">"""Preprocess ProSnow dataset images.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> data : `numpy.ndarray`</span>
-<span class="sd"> The sample input data.</span>
-<span class="sd"> gt : `numpy.ndarray`</span>
-<span class="sd"> The sample ground truth.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> data : `numpy.ndarray`</span>
-<span class="sd"> The preprocessed input data.</span>
-<span class="sd"> gt : `numpy.ndarray`</span>
-<span class="sd"> The preprocessed ground truth data.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># if the preprocessing is not done externally, implement it here</span>
- <span class="k">return</span> <span class="n">data</span><span class="p">,</span> <span class="n">gt</span></div>
-
-<div class="viewcode-block" id="ProSnowDataset.parse_scene_id"><a class="viewcode-back" href="../../source/core.html#core.dataset.ProSnowDataset.parse_scene_id">[docs]</a> <span class="k">def</span> <span class="nf">parse_scene_id</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">scene_id</span><span class="p">):</span>
- <span class="sd">"""Parse ProSnow scene identifiers (Sentinel 2).</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> scene_id : `str`</span>
-<span class="sd"> A scene identifier.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> scene : `dict` or `None`</span>
-<span class="sd"> A dictionary containing scene metadata. If `None`, ``scene_id`` is</span>
-<span class="sd"> not a valid Sentinel-2 scene identifier.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="n">parse_sentinel2_scene</span><span class="p">(</span><span class="n">scene_id</span><span class="p">)</span></div></div>
-
-
-<div class="viewcode-block" id="ProSnowGarmisch"><a class="viewcode-back" href="../../source/core.html#core.dataset.ProSnowGarmisch">[docs]</a><span class="k">class</span> <span class="nc">ProSnowGarmisch</span><span class="p">(</span><span class="n">ProSnowDataset</span><span class="p">):</span>
- <span class="sa">r</span><span class="sd">"""Class for the ProSnow Garmisch dataset.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> root_dir : `str`</span>
-<span class="sd"> The root directory, path to the dataset.</span>
-<span class="sd"> use_bands : `list` [`str`], optional</span>
-<span class="sd"> A list of the spectral bands to use. The default is [].</span>
-<span class="sd"> tile_size : `int` or `None`, optional</span>
-<span class="sd"> The size of the tiles. If not `None`, each scene is divided into square</span>
-<span class="sd"> tiles of shape (tile_size, tile_size). The default is None.</span>
-<span class="sd"> pad : `bool`, optional</span>
-<span class="sd"> Whether to center pad the input image. Set ``pad`` = True, if the</span>
-<span class="sd"> images are not evenly divisible by the ``tile_size``. The image data is</span>
-<span class="sd"> padded with a constant padding value of zero. For each image, the</span>
-<span class="sd"> corresponding ground truth image is padded with a "no data" label.</span>
-<span class="sd"> The default is False.</span>
-<span class="sd"> gt_pattern : `str`, optional</span>
-<span class="sd"> A regural expression to match the ground truth naming convention. All</span>
-<span class="sd"> directories and subdirectories in ``root_dir`` are searched for files</span>
-<span class="sd"> matching ``gt_pattern``. The default is '(.*)gt\\.tif'.</span>
-<span class="sd"> sort : `bool`, optional</span>
-<span class="sd"> Whether to chronologically sort the samples. Useful for time series</span>
-<span class="sd"> data. The default is False.</span>
-<span class="sd"> seed : `int`, optional</span>
-<span class="sd"> The random seed. Used to split the dataset into training, validation</span>
-<span class="sd"> and test set. Useful for reproducibility. The default is 0.</span>
-<span class="sd"> transforms : `list` [`pysegcnn.core.split.Augment`], optional</span>
-<span class="sd"> List of `pysegcnn.core.split.Augment` instances. Each item in</span>
-<span class="sd"> ``transforms`` generates a distinct transformed version of the dataset.</span>
-<span class="sd"> The total dataset is composed of the original untransformed dataset</span>
-<span class="sd"> together with each transformed version of it.</span>
-<span class="sd"> If ``transforms`` = [], only the original dataset is used.</span>
-<span class="sd"> The default is [].</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">root_dir</span><span class="p">,</span> <span class="n">use_bands</span><span class="o">=</span><span class="p">[],</span> <span class="n">tile_size</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">pad</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span>
- <span class="n">gt_pattern</span><span class="o">=</span><span class="s1">'(.*)gt</span><span class="se">\\</span><span class="s1">.tif'</span><span class="p">,</span> <span class="n">sort</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">0</span><span class="p">,</span> <span class="n">transforms</span><span class="o">=</span><span class="p">[]):</span>
- <span class="c1"># initialize super class StandardEoDataset</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">root_dir</span><span class="p">,</span> <span class="n">use_bands</span><span class="p">,</span> <span class="n">tile_size</span><span class="p">,</span> <span class="n">pad</span><span class="p">,</span> <span class="n">gt_pattern</span><span class="p">,</span>
- <span class="n">sort</span><span class="p">,</span> <span class="n">seed</span><span class="p">,</span> <span class="n">transforms</span><span class="p">)</span>
-
-<div class="viewcode-block" id="ProSnowGarmisch.get_size"><a class="viewcode-back" href="../../source/core.html#core.dataset.ProSnowGarmisch.get_size">[docs]</a> <span class="k">def</span> <span class="nf">get_size</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Image size of the ProSnow Garmisch dataset.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> size : `tuple`</span>
-<span class="sd"> The image size (height, width).</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="p">(</span><span class="mi">615</span><span class="p">,</span> <span class="mi">543</span><span class="p">)</span></div></div>
-
-
-<div class="viewcode-block" id="ProSnowObergurgl"><a class="viewcode-back" href="../../source/core.html#core.dataset.ProSnowObergurgl">[docs]</a><span class="k">class</span> <span class="nc">ProSnowObergurgl</span><span class="p">(</span><span class="n">ProSnowDataset</span><span class="p">):</span>
- <span class="sa">r</span><span class="sd">"""Class for the ProSnow Obergurgl dataset.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> root_dir : `str`</span>
-<span class="sd"> The root directory, path to the dataset.</span>
-<span class="sd"> use_bands : `list` [`str`], optional</span>
-<span class="sd"> A list of the spectral bands to use. The default is [].</span>
-<span class="sd"> tile_size : `int` or `None`, optional</span>
-<span class="sd"> The size of the tiles. If not `None`, each scene is divided into square</span>
-<span class="sd"> tiles of shape (tile_size, tile_size). The default is None.</span>
-<span class="sd"> pad : `bool`, optional</span>
-<span class="sd"> Whether to center pad the input image. Set ``pad`` = True, if the</span>
-<span class="sd"> images are not evenly divisible by the ``tile_size``. The image data is</span>
-<span class="sd"> padded with a constant padding value of zero. For each image, the</span>
-<span class="sd"> corresponding ground truth image is padded with a "no data" label.</span>
-<span class="sd"> The default is False.</span>
-<span class="sd"> gt_pattern : `str`, optional</span>
-<span class="sd"> A regural expression to match the ground truth naming convention. All</span>
-<span class="sd"> directories and subdirectories in ``root_dir`` are searched for files</span>
-<span class="sd"> matching ``gt_pattern``. The default is '(.*)gt\\.tif'.</span>
-<span class="sd"> sort : `bool`, optional</span>
-<span class="sd"> Whether to chronologically sort the samples. Useful for time series</span>
-<span class="sd"> data. The default is False.</span>
-<span class="sd"> seed : `int`, optional</span>
-<span class="sd"> The random seed. Used to split the dataset into training, validation</span>
-<span class="sd"> and test set. Useful for reproducibility. The default is 0.</span>
-<span class="sd"> transforms : `list` [`pysegcnn.core.split.Augment`], optional</span>
-<span class="sd"> List of `pysegcnn.core.split.Augment` instances. Each item in</span>
-<span class="sd"> ``transforms`` generates a distinct transformed version of the dataset.</span>
-<span class="sd"> The total dataset is composed of the original untransformed dataset</span>
-<span class="sd"> together with each transformed version of it.</span>
-<span class="sd"> If ``transforms`` = [], only the original dataset is used.</span>
-<span class="sd"> The default is [].</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">root_dir</span><span class="p">,</span> <span class="n">use_bands</span><span class="o">=</span><span class="p">[],</span> <span class="n">tile_size</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">pad</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span>
- <span class="n">gt_pattern</span><span class="o">=</span><span class="s1">'(.*)gt</span><span class="se">\\</span><span class="s1">.tif'</span><span class="p">,</span> <span class="n">sort</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">0</span><span class="p">,</span> <span class="n">transforms</span><span class="o">=</span><span class="p">[]):</span>
- <span class="c1"># initialize super class StandardEoDataset</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">root_dir</span><span class="p">,</span> <span class="n">use_bands</span><span class="p">,</span> <span class="n">tile_size</span><span class="p">,</span> <span class="n">pad</span><span class="p">,</span> <span class="n">gt_pattern</span><span class="p">,</span>
- <span class="n">sort</span><span class="p">,</span> <span class="n">seed</span><span class="p">,</span> <span class="n">transforms</span><span class="p">)</span>
-
-<div class="viewcode-block" id="ProSnowObergurgl.get_size"><a class="viewcode-back" href="../../source/core.html#core.dataset.ProSnowObergurgl.get_size">[docs]</a> <span class="k">def</span> <span class="nf">get_size</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Image size of the ProSnow Obergurgl dataset.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> size : `tuple`</span>
-<span class="sd"> The image size (height, width).</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="p">(</span><span class="mi">310</span><span class="p">,</span> <span class="mi">270</span><span class="p">)</span></div></div>
-
-
-<div class="viewcode-block" id="Cloud95Dataset"><a class="viewcode-back" href="../../source/core.html#core.dataset.Cloud95Dataset">[docs]</a><span class="k">class</span> <span class="nc">Cloud95Dataset</span><span class="p">(</span><span class="n">ImageDataset</span><span class="p">):</span>
- <span class="sa">r</span><span class="sd">"""Class for the `Cloud-95`_ dataset by `Mohajerani et al. (2020)`_.</span>
-
-<span class="sd"> .. _Cloud95:</span>
-<span class="sd"> https://github.com/SorourMo/95-Cloud-An-Extension-to-38-Cloud-Dataset</span>
-<span class="sd"> .. _`Mohajerani et al. (2020)`:</span>
-<span class="sd"> https://arxiv.org/abs/2001.08768</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> root_dir : `str`</span>
-<span class="sd"> The root directory, path to the dataset.</span>
-<span class="sd"> use_bands : `list` [`str`], optional</span>
-<span class="sd"> A list of the spectral bands to use. The default is [].</span>
-<span class="sd"> tile_size : `int` or `None`, optional</span>
-<span class="sd"> The size of the tiles. If not `None`, each scene is divided into square</span>
-<span class="sd"> tiles of shape (tile_size, tile_size). The default is None.</span>
-<span class="sd"> pad : `bool`, optional</span>
-<span class="sd"> Whether to center pad the input image. Set ``pad`` = True, if the</span>
-<span class="sd"> images are not evenly divisible by the ``tile_size``. The image data is</span>
-<span class="sd"> padded with a constant padding value of zero. For each image, the</span>
-<span class="sd"> corresponding ground truth image is padded with a "no data" label.</span>
-<span class="sd"> The default is False.</span>
-<span class="sd"> gt_pattern : `str`, optional</span>
-<span class="sd"> A regural expression to match the ground truth naming convention. All</span>
-<span class="sd"> directories and subdirectories in ``root_dir`` are searched for files</span>
-<span class="sd"> matching ``gt_pattern``. The default is '(.*)gt\\.tif'.</span>
-<span class="sd"> sort : `bool`, optional</span>
-<span class="sd"> Whether to chronologically sort the samples. Useful for time series</span>
-<span class="sd"> data. The default is False.</span>
-<span class="sd"> seed : `int`, optional</span>
-<span class="sd"> The random seed. Used to split the dataset into training, validation</span>
-<span class="sd"> and test set. Useful for reproducibility. The default is 0.</span>
-<span class="sd"> transforms : `list` [`pysegcnn.core.split.Augment`], optional</span>
-<span class="sd"> List of `pysegcnn.core.split.Augment` instances. Each item in</span>
-<span class="sd"> ``transforms`` generates a distinct transformed version of the dataset.</span>
-<span class="sd"> The total dataset is composed of the original untransformed dataset</span>
-<span class="sd"> together with each transformed version of it.</span>
-<span class="sd"> If ``transforms`` = [], only the original dataset is used.</span>
-<span class="sd"> The default is [].</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">root_dir</span><span class="p">,</span> <span class="n">use_bands</span><span class="o">=</span><span class="p">[],</span> <span class="n">tile_size</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">pad</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span>
- <span class="n">gt_pattern</span><span class="o">=</span><span class="s1">'(.*)gt</span><span class="se">\\</span><span class="s1">.tif'</span><span class="p">,</span> <span class="n">sort</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">0</span><span class="p">,</span> <span class="n">transforms</span><span class="o">=</span><span class="p">[]):</span>
- <span class="c1"># initialize super class StandardEoDataset</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">root_dir</span><span class="p">,</span> <span class="n">use_bands</span><span class="p">,</span> <span class="n">tile_size</span><span class="p">,</span> <span class="n">pad</span><span class="p">,</span> <span class="n">gt_pattern</span><span class="p">,</span>
- <span class="n">sort</span><span class="p">,</span> <span class="n">seed</span><span class="p">,</span> <span class="n">transforms</span><span class="p">)</span>
-
- <span class="c1"># the csv file containing the names of the informative patches</span>
- <span class="c1"># patches with more than 80% black pixels, i.e. patches resulting from</span>
- <span class="c1"># the black margins around a Landsat 8 scene are excluded</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">exclude</span> <span class="o">=</span> <span class="s1">'training_patches_95-cloud_nonempty.csv'</span>
-
-<div class="viewcode-block" id="Cloud95Dataset.get_size"><a class="viewcode-back" href="../../source/core.html#core.dataset.Cloud95Dataset.get_size">[docs]</a> <span class="k">def</span> <span class="nf">get_size</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Image size of the Cloud-95 dataset.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> size : `tuple`</span>
-<span class="sd"> The image size (height, width).</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="p">(</span><span class="mi">384</span><span class="p">,</span> <span class="mi">384</span><span class="p">)</span></div>
-
-<div class="viewcode-block" id="Cloud95Dataset.get_sensor"><a class="viewcode-back" href="../../source/core.html#core.dataset.Cloud95Dataset.get_sensor">[docs]</a> <span class="k">def</span> <span class="nf">get_sensor</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Landsat 8 bands of the Cloud-95 dataset.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> sensor : `enum.Enum`</span>
-<span class="sd"> An enumeration of the bands of the sensor.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="n">Landsat8</span></div>
-
-<div class="viewcode-block" id="Cloud95Dataset.get_labels"><a class="viewcode-back" href="../../source/core.html#core.dataset.Cloud95Dataset.get_labels">[docs]</a> <span class="k">def</span> <span class="nf">get_labels</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Class labels of the Cloud-95 dataset.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> labels : `enum.Enum`</span>
-<span class="sd"> The class labels.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="n">Cloud95Labels</span></div>
-
-<div class="viewcode-block" id="Cloud95Dataset.preprocess"><a class="viewcode-back" href="../../source/core.html#core.dataset.Cloud95Dataset.preprocess">[docs]</a> <span class="k">def</span> <span class="nf">preprocess</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">data</span><span class="p">,</span> <span class="n">gt</span><span class="p">):</span>
- <span class="sd">"""Preprocess Cloud-95 dataset images.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> data : `numpy.ndarray`</span>
-<span class="sd"> The sample input data.</span>
-<span class="sd"> gt : `numpy.ndarray`</span>
-<span class="sd"> The sample ground truth.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> data : `numpy.ndarray`</span>
-<span class="sd"> The preprocessed input data.</span>
-<span class="sd"> gt : `numpy.ndarray`</span>
-<span class="sd"> The preprocessed ground truth data.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># normalize the data</span>
- <span class="c1"># here, we use the normalization of the authors of Cloud-95, i.e.</span>
- <span class="c1"># Mohajerani and Saeedi (2019, 2020)</span>
- <span class="n">data</span> <span class="o">/=</span> <span class="mi">65535</span>
- <span class="n">gt</span><span class="p">[</span><span class="n">gt</span> <span class="o">!=</span> <span class="bp">self</span><span class="o">.</span><span class="n">cval</span><span class="p">]</span> <span class="o">/=</span> <span class="mi">255</span>
- <span class="k">return</span> <span class="n">data</span><span class="p">,</span> <span class="n">gt</span></div>
-
-<div class="viewcode-block" id="Cloud95Dataset.parse_scene_id"><a class="viewcode-back" href="../../source/core.html#core.dataset.Cloud95Dataset.parse_scene_id">[docs]</a> <span class="k">def</span> <span class="nf">parse_scene_id</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">scene_id</span><span class="p">):</span>
- <span class="sd">"""Parse Sparcs scene identifiers (Landsat 8).</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> scene_id : `str`</span>
-<span class="sd"> A scene identifier.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> scene : `dict` or `None`</span>
-<span class="sd"> A dictionary containing scene metadata. If `None`, ``scene_id`` is</span>
-<span class="sd"> not a valid Landsat scene identifier.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="n">parse_landsat_scene</span><span class="p">(</span><span class="n">scene_id</span><span class="p">)</span></div>
-
-<div class="viewcode-block" id="Cloud95Dataset.compose_scenes"><a class="viewcode-back" href="../../source/core.html#core.dataset.Cloud95Dataset.compose_scenes">[docs]</a> <span class="k">def</span> <span class="nf">compose_scenes</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Build the list of samples of the dataset.</span>
-
-<span class="sd"> Each sample is represented by a dictionary.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> scenes : `list` [`dict`]</span>
-<span class="sd"> Each item in ``scenes`` is a `dict` with keys:</span>
-<span class="sd"> ``'band_name_1'``</span>
-<span class="sd"> Path to the file of band_1.</span>
-<span class="sd"> ``'band_name_2'``</span>
-<span class="sd"> Path to the file of band_2.</span>
-<span class="sd"> ``'band_name_n'``</span>
-<span class="sd"> Path to the file of band_n.</span>
-<span class="sd"> ``'gt'``</span>
-<span class="sd"> Path to the ground truth file.</span>
-<span class="sd"> ``'date'``</span>
-<span class="sd"> The date of the sample.</span>
-<span class="sd"> ``'tile'``</span>
-<span class="sd"> The tile id of the sample.</span>
-<span class="sd"> ``'transform'``</span>
-<span class="sd"> The transformation to apply.</span>
-<span class="sd"> ``'id'``</span>
-<span class="sd"> The scene identifier.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># whether to exclude patches with more than 80% black pixels</span>
- <span class="n">ipatches</span> <span class="o">=</span> <span class="p">[]</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">exclude</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
- <span class="k">with</span> <span class="nb">open</span><span class="p">(</span><span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">root</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">exclude</span><span class="p">),</span> <span class="n">newline</span><span class="o">=</span><span class="s1">''</span><span class="p">)</span> <span class="k">as</span> <span class="n">f</span><span class="p">:</span>
- <span class="n">reader</span> <span class="o">=</span> <span class="n">csv</span><span class="o">.</span><span class="n">reader</span><span class="p">(</span><span class="n">f</span><span class="p">)</span>
- <span class="c1"># list of informative patches</span>
- <span class="n">ipatches</span> <span class="o">=</span> <span class="nb">list</span><span class="p">(</span><span class="n">itertools</span><span class="o">.</span><span class="n">chain</span><span class="o">.</span><span class="n">from_iterable</span><span class="p">(</span><span class="n">reader</span><span class="p">))</span>
-
- <span class="c1"># get the names of the directories containing the tif files of</span>
- <span class="c1"># the bands of interest</span>
- <span class="n">band_dirs</span> <span class="o">=</span> <span class="p">{}</span>
- <span class="k">for</span> <span class="n">dirpath</span><span class="p">,</span> <span class="n">dirname</span><span class="p">,</span> <span class="n">files</span> <span class="ow">in</span> <span class="n">os</span><span class="o">.</span><span class="n">walk</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">root</span><span class="p">):</span>
- <span class="c1"># check if the current directory path includes the name of a band</span>
- <span class="c1"># or the name of the ground truth mask</span>
- <span class="n">cband</span> <span class="o">=</span> <span class="p">[</span><span class="n">band</span> <span class="k">for</span> <span class="n">band</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">use_bands</span> <span class="o">+</span> <span class="p">[</span><span class="s1">'gt'</span><span class="p">]</span> <span class="k">if</span>
- <span class="n">dirpath</span><span class="o">.</span><span class="n">endswith</span><span class="p">(</span><span class="n">band</span><span class="p">)</span> <span class="ow">and</span> <span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">isdir</span><span class="p">(</span><span class="n">dirpath</span><span class="p">)]</span>
-
- <span class="c1"># add path to current band files to dictionary</span>
- <span class="k">if</span> <span class="n">cband</span><span class="p">:</span>
- <span class="n">band_dirs</span><span class="p">[</span><span class="n">cband</span><span class="o">.</span><span class="n">pop</span><span class="p">()]</span> <span class="o">=</span> <span class="n">dirpath</span>
-
- <span class="c1"># create empty list to store all patches to</span>
- <span class="n">scenes</span> <span class="o">=</span> <span class="p">[]</span>
-
- <span class="c1"># iterate over all the patches of the following band</span>
- <span class="n">biter</span> <span class="o">=</span> <span class="p">[</span><span class="o">*</span><span class="n">band_dirs</span><span class="o">.</span><span class="n">keys</span><span class="p">()][</span><span class="mi">0</span><span class="p">]</span>
- <span class="k">for</span> <span class="n">file</span> <span class="ow">in</span> <span class="n">os</span><span class="o">.</span><span class="n">listdir</span><span class="p">(</span><span class="n">band_dirs</span><span class="p">[</span><span class="n">biter</span><span class="p">]):</span>
-
- <span class="c1"># get name of the current patch</span>
- <span class="n">patchname</span> <span class="o">=</span> <span class="n">file</span><span class="o">.</span><span class="n">split</span><span class="p">(</span><span class="s1">'.'</span><span class="p">)[</span><span class="mi">0</span><span class="p">]</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="n">biter</span> <span class="o">+</span> <span class="s1">'_'</span><span class="p">,</span> <span class="s1">''</span><span class="p">)</span>
-
- <span class="c1"># get the date of the current scene</span>
- <span class="n">scene_meta</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">parse_scene_id</span><span class="p">(</span><span class="n">patchname</span><span class="p">)</span>
-
- <span class="c1"># check whether the current file is an informative patch</span>
- <span class="k">if</span> <span class="n">ipatches</span> <span class="ow">and</span> <span class="n">patchname</span> <span class="ow">not</span> <span class="ow">in</span> <span class="n">ipatches</span><span class="p">:</span>
- <span class="k">continue</span>
-
- <span class="c1"># iterate over the tiles</span>
- <span class="k">for</span> <span class="n">tile</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">tiles</span><span class="p">):</span>
-
- <span class="c1"># iterate over the transformations to apply</span>
- <span class="k">for</span> <span class="n">transf</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">transforms</span><span class="p">:</span>
-
- <span class="c1"># initialize dictionary to store bands of current patch</span>
- <span class="n">scene</span> <span class="o">=</span> <span class="p">{}</span>
-
- <span class="c1"># iterate over the bands of interest</span>
- <span class="k">for</span> <span class="n">band</span> <span class="ow">in</span> <span class="n">band_dirs</span><span class="o">.</span><span class="n">keys</span><span class="p">():</span>
- <span class="c1"># save path to current band tif file to dictionary</span>
- <span class="n">scene</span><span class="p">[</span><span class="n">band</span><span class="p">]</span> <span class="o">=</span> <span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">band_dirs</span><span class="p">[</span><span class="n">band</span><span class="p">],</span>
- <span class="n">file</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="n">biter</span><span class="p">,</span> <span class="n">band</span><span class="p">))</span>
-
- <span class="c1"># the name of the scene the patch was extracted from</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'id'</span><span class="p">]</span> <span class="o">=</span> <span class="n">scene_meta</span><span class="p">[</span><span class="s1">'id'</span><span class="p">]</span>
-
- <span class="c1"># store tile number</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'tile'</span><span class="p">]</span> <span class="o">=</span> <span class="n">tile</span>
-
- <span class="c1"># store date</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'date'</span><span class="p">]</span> <span class="o">=</span> <span class="n">scene_meta</span><span class="p">[</span><span class="s1">'date'</span><span class="p">]</span>
-
- <span class="c1"># store optional transformation</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'transform'</span><span class="p">]</span> <span class="o">=</span> <span class="n">transf</span>
-
- <span class="c1"># append patch to list of all patches</span>
- <span class="n">scenes</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">scene</span><span class="p">)</span>
-
- <span class="c1"># sort list of scenes in chronological order</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">sort</span><span class="p">:</span>
- <span class="n">scenes</span><span class="o">.</span><span class="n">sort</span><span class="p">(</span><span class="n">key</span><span class="o">=</span><span class="k">lambda</span> <span class="n">k</span><span class="p">:</span> <span class="n">k</span><span class="p">[</span><span class="s1">'date'</span><span class="p">])</span>
-
- <span class="k">return</span> <span class="n">scenes</span></div></div>
-
-
-<div class="viewcode-block" id="SupportedDatasets"><a class="viewcode-back" href="../../source/core.html#core.dataset.SupportedDatasets">[docs]</a><span class="k">class</span> <span class="nc">SupportedDatasets</span><span class="p">(</span><span class="n">enum</span><span class="o">.</span><span class="n">Enum</span><span class="p">):</span>
- <span class="sd">"""Names and corresponding classes of the implemented datasets."""</span>
-
- <span class="n">Sparcs</span> <span class="o">=</span> <span class="n">SparcsDataset</span>
- <span class="n">Cloud95</span> <span class="o">=</span> <span class="n">Cloud95Dataset</span>
- <span class="n">Garmisch</span> <span class="o">=</span> <span class="n">ProSnowGarmisch</span>
- <span class="n">Obergurgl</span> <span class="o">=</span> <span class="n">ProSnowObergurgl</span></div>
-</pre></div>
-
- </div>
-
- </div>
- <footer>
-
-
- <hr/>
-
- <div role="contentinfo">
- <p>
-
- © Copyright 2020, Daniel Frisinghelli
-
- </p>
- </div>
-
-
-
- Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a
-
- <a href="https://github.com/rtfd/sphinx_rtd_theme">theme</a>
-
- provided by <a href="https://readthedocs.org">Read the Docs</a>.
-
-</footer>
-
- </div>
- </div>
-
- </section>
-
- </div>
-
-
- <script type="text/javascript">
- jQuery(function () {
- SphinxRtdTheme.Navigation.enable(true);
- });
- </script>
-
-
-
-
-
-
-</body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/_modules/core/graphics.html b/docs/_build/html/_modules/core/graphics.html
deleted file mode 100644
index 4ed801e9a6fad5fb01791535cea76224e5445f75..0000000000000000000000000000000000000000
--- a/docs/_build/html/_modules/core/graphics.html
+++ /dev/null
@@ -1,592 +0,0 @@
-
-
-<!DOCTYPE html>
-<html class="writer-html5" lang="en" >
-<head>
- <meta charset="utf-8">
-
- <meta name="viewport" content="width=device-width, initial-scale=1.0">
-
- <title>core.graphics — PySegCNN documentation</title>
-
-
-
- <link rel="stylesheet" href="../../_static/css/theme.css" type="text/css" />
- <link rel="stylesheet" href="../../_static/pygments.css" type="text/css" />
-
-
-
-
-
-
-
- <!--[if lt IE 9]>
- <script src="../../_static/js/html5shiv.min.js"></script>
- <![endif]-->
-
-
- <script type="text/javascript" id="documentation_options" data-url_root="../../" src="../../_static/documentation_options.js"></script>
- <script src="../../_static/jquery.js"></script>
- <script src="../../_static/underscore.js"></script>
- <script src="../../_static/doctools.js"></script>
- <script src="../../_static/language_data.js"></script>
-
- <script type="text/javascript" src="../../_static/js/theme.js"></script>
-
-
- <link rel="index" title="Index" href="../../genindex.html" />
- <link rel="search" title="Search" href="../../search.html" />
-</head>
-
-<body class="wy-body-for-nav">
-
-
- <div class="wy-grid-for-nav">
-
- <nav data-toggle="wy-nav-shift" class="wy-nav-side">
- <div class="wy-side-scroll">
- <div class="wy-side-nav-search" >
-
-
-
- <a href="../../index.html" class="icon icon-home" alt="Documentation Home"> PySegCNN
-
-
-
- </a>
-
-
-
-
-
-
-
-<div role="search">
- <form id="rtd-search-form" class="wy-form" action="../../search.html" method="get">
- <input type="text" name="q" placeholder="Search docs" />
- <input type="hidden" name="check_keywords" value="yes" />
- <input type="hidden" name="area" value="default" />
- </form>
-</div>
-
-
- </div>
-
-
- <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
-
-
-
-
-
-
- <ul>
-<li class="toctree-l1"><a class="reference internal" href="../../source/requirements.html">Requirements</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../../source/installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../../source/api.html">API Reference</a></li>
-</ul>
-
-
-
- </div>
-
- </div>
- </nav>
-
- <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
-
-
- <nav class="wy-nav-top" aria-label="top navigation">
-
- <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
- <a href="../../index.html">PySegCNN</a>
-
- </nav>
-
-
- <div class="wy-nav-content">
-
- <div class="rst-content">
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-<div role="navigation" aria-label="breadcrumbs navigation">
-
- <ul class="wy-breadcrumbs">
-
- <li><a href="../../index.html" class="icon icon-home"></a> »</li>
-
- <li><a href="../index.html">Module code</a> »</li>
-
- <li>core.graphics</li>
-
-
- <li class="wy-breadcrumbs-aside">
-
- </li>
-
- </ul>
-
-
- <hr/>
-</div>
- <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
- <div itemprop="articleBody">
-
- <h1>Source code for core.graphics</h1><div class="highlight"><pre>
-<span></span><span class="sd">"""Functions to plot multispectral image data and model output.</span>
-
-<span class="sd">License</span>
-<span class="sd">-------</span>
-
-<span class="sd"> Copyright (c) 2020 Daniel Frisinghelli</span>
-
-<span class="sd"> This source code is licensed under the GNU General Public License v3.</span>
-
-<span class="sd"> See the LICENSE file in the repository's root directory.</span>
-
-<span class="sd">"""</span>
-
-<span class="c1"># !/usr/bin/env python</span>
-<span class="c1"># -*- coding: utf-8 -*-</span>
-
-<span class="c1"># builtins</span>
-<span class="kn">import</span> <span class="nn">os</span>
-<span class="kn">import</span> <span class="nn">itertools</span>
-
-<span class="c1"># externals</span>
-<span class="kn">import</span> <span class="nn">numpy</span> <span class="k">as</span> <span class="nn">np</span>
-<span class="kn">import</span> <span class="nn">torch</span>
-<span class="kn">import</span> <span class="nn">matplotlib.pyplot</span> <span class="k">as</span> <span class="nn">plt</span>
-<span class="kn">import</span> <span class="nn">matplotlib.patches</span> <span class="k">as</span> <span class="nn">mpatches</span>
-<span class="kn">import</span> <span class="nn">matplotlib.lines</span> <span class="k">as</span> <span class="nn">mlines</span>
-<span class="kn">from</span> <span class="nn">matplotlib.colors</span> <span class="kn">import</span> <span class="n">ListedColormap</span><span class="p">,</span> <span class="n">BoundaryNorm</span>
-<span class="kn">from</span> <span class="nn">matplotlib</span> <span class="kn">import</span> <span class="n">cm</span> <span class="k">as</span> <span class="n">colormap</span>
-
-<span class="c1"># locals</span>
-<span class="kn">from</span> <span class="nn">pysegcnn.core.trainer</span> <span class="kn">import</span> <span class="n">accuracy_function</span>
-<span class="kn">from</span> <span class="nn">pysegcnn.main.config</span> <span class="kn">import</span> <span class="n">HERE</span>
-
-
-<div class="viewcode-block" id="contrast_stretching"><a class="viewcode-back" href="../../source/core.html#core.graphics.contrast_stretching">[docs]</a><span class="k">def</span> <span class="nf">contrast_stretching</span><span class="p">(</span><span class="n">image</span><span class="p">,</span> <span class="n">alpha</span><span class="o">=</span><span class="mi">5</span><span class="p">):</span>
- <span class="sd">"""Apply percentile stretching to an image to increase constrast.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> image : `numpy.ndarray`</span>
-<span class="sd"> the input image.</span>
-<span class="sd"> alpha : `int`, optional</span>
-<span class="sd"> The level of the percentiles. The default is 5.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> norm : `numpy.ndarray`</span>
-<span class="sd"> the stretched image.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># compute upper and lower percentiles defining the range of the stretch</span>
- <span class="n">inf</span><span class="p">,</span> <span class="n">sup</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">percentile</span><span class="p">(</span><span class="n">image</span><span class="p">,</span> <span class="p">(</span><span class="n">alpha</span><span class="p">,</span> <span class="mi">100</span> <span class="o">-</span> <span class="n">alpha</span><span class="p">))</span>
-
- <span class="c1"># normalize image intensity distribution to</span>
- <span class="c1"># (alpha, 100 - alpha) percentiles</span>
- <span class="n">norm</span> <span class="o">=</span> <span class="p">((</span><span class="n">image</span> <span class="o">-</span> <span class="n">inf</span><span class="p">)</span> <span class="o">*</span> <span class="p">(</span><span class="n">image</span><span class="o">.</span><span class="n">max</span><span class="p">()</span> <span class="o">-</span> <span class="n">image</span><span class="o">.</span><span class="n">min</span><span class="p">())</span> <span class="o">/</span>
- <span class="p">(</span><span class="n">sup</span> <span class="o">-</span> <span class="n">inf</span><span class="p">))</span> <span class="o">+</span> <span class="n">image</span><span class="o">.</span><span class="n">min</span><span class="p">()</span>
-
- <span class="c1"># clip: values < min = min, values > max = max</span>
- <span class="n">norm</span><span class="p">[</span><span class="n">norm</span> <span class="o"><=</span> <span class="n">image</span><span class="o">.</span><span class="n">min</span><span class="p">()]</span> <span class="o">=</span> <span class="n">image</span><span class="o">.</span><span class="n">min</span><span class="p">()</span>
- <span class="n">norm</span><span class="p">[</span><span class="n">norm</span> <span class="o">>=</span> <span class="n">image</span><span class="o">.</span><span class="n">max</span><span class="p">()]</span> <span class="o">=</span> <span class="n">image</span><span class="o">.</span><span class="n">max</span><span class="p">()</span>
-
- <span class="k">return</span> <span class="n">norm</span></div>
-
-
-<div class="viewcode-block" id="running_mean"><a class="viewcode-back" href="../../source/core.html#core.graphics.running_mean">[docs]</a><span class="k">def</span> <span class="nf">running_mean</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">w</span><span class="p">):</span>
- <span class="sd">"""Compute a running mean of the input sequence.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> x : array_like</span>
-<span class="sd"> The sequence to compute a running mean on.</span>
-<span class="sd"> w : `int`</span>
-<span class="sd"> The window length of the running mean.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> rm : `numpy.ndarray`</span>
-<span class="sd"> The running mean of the sequence ``x``.</span>
-
-<span class="sd"> """</span>
- <span class="n">cumsum</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">cumsum</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">insert</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">0</span><span class="p">))</span>
- <span class="k">return</span> <span class="p">(</span><span class="n">cumsum</span><span class="p">[</span><span class="n">w</span><span class="p">:]</span> <span class="o">-</span> <span class="n">cumsum</span><span class="p">[:</span><span class="o">-</span><span class="n">w</span><span class="p">])</span> <span class="o">/</span> <span class="n">w</span></div>
-
-
-<div class="viewcode-block" id="plot_sample"><a class="viewcode-back" href="../../source/core.html#core.graphics.plot_sample">[docs]</a><span class="k">def</span> <span class="nf">plot_sample</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">use_bands</span><span class="p">,</span> <span class="n">labels</span><span class="p">,</span> <span class="n">y</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">y_pred</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">figsize</span><span class="o">=</span><span class="p">(</span><span class="mi">10</span><span class="p">,</span> <span class="mi">10</span><span class="p">),</span>
- <span class="n">bands</span><span class="o">=</span><span class="p">[</span><span class="s1">'nir'</span><span class="p">,</span> <span class="s1">'red'</span><span class="p">,</span> <span class="s1">'green'</span><span class="p">],</span> <span class="n">state</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span>
- <span class="n">outpath</span><span class="o">=</span><span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">HERE</span><span class="p">,</span> <span class="s1">'_samples/'</span><span class="p">),</span> <span class="n">alpha</span><span class="o">=</span><span class="mi">0</span><span class="p">):</span>
- <span class="sd">"""Plot false color composite (FCC), ground truth and model prediction.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> x : `numpy.ndarray` or `torch.Tensor`, (b, h, w)</span>
-<span class="sd"> Array containing the raw data of the tile, shape=(bands, height, width)</span>
-<span class="sd"> use_bands : `list` of `str`</span>
-<span class="sd"> List describing the order of the bands in ``x``.</span>
-<span class="sd"> labels : `dict` [`int`, `dict`]</span>
-<span class="sd"> The label dictionary. The keys are the values of the class labels</span>
-<span class="sd"> in the ground truth ``y``. Each nested `dict` should have keys:</span>
-<span class="sd"> ``'color'``</span>
-<span class="sd"> A named color (`str`).</span>
-<span class="sd"> ``'label'``</span>
-<span class="sd"> The name of the class label (`str`).</span>
-<span class="sd"> y : `numpy.ndarray` or `torch.Tensor` or `None`, optional</span>
-<span class="sd"> Array containing the ground truth of tile ``x``, shape=(height, width).</span>
-<span class="sd"> The default is None.</span>
-<span class="sd"> y_pred : `numpy.ndarray` or `torch.Tensor` or `None`, optional</span>
-<span class="sd"> Array containing the prediction for tile ``x``, shape=(height, width).</span>
-<span class="sd"> The default is None.</span>
-<span class="sd"> figsize : `tuple`, optional</span>
-<span class="sd"> The figure size in centimeters. The default is (10, 10).</span>
-<span class="sd"> bands : `list` [`str`], optional</span>
-<span class="sd"> The bands to build the FCC. The default is ['nir', 'red', 'green'].</span>
-<span class="sd"> state : `str` or `None`, optional</span>
-<span class="sd"> Filename to save the plot to. ``state`` should be an existing model</span>
-<span class="sd"> state file ending with '.pt'. The default is None, i.e. plot is not</span>
-<span class="sd"> saved to disk.</span>
-<span class="sd"> outpath : `str` or `pathlib.Path`, optional</span>
-<span class="sd"> Output path. The default is 'pysegcnn/main/_samples'.</span>
-<span class="sd"> alpha : `int`, optional</span>
-<span class="sd"> The level of the percentiles to increase constrast in the FCC.</span>
-<span class="sd"> The default is 0, i.e. no stretching.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> fig : `matplotlib.figure.Figure`</span>
-<span class="sd"> The figure handle.</span>
-<span class="sd"> ax : `numpy.ndarray` [`matplotlib.axes._subplots.AxesSubplot`]</span>
-<span class="sd"> An array of the axes handles.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># check whether to apply constrast stretching</span>
- <span class="n">rgb</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">dstack</span><span class="p">([</span><span class="n">contrast_stretching</span><span class="p">(</span><span class="n">x</span><span class="p">[</span><span class="n">use_bands</span><span class="o">.</span><span class="n">index</span><span class="p">(</span><span class="n">band</span><span class="p">)],</span> <span class="n">alpha</span><span class="p">)</span>
- <span class="k">for</span> <span class="n">band</span> <span class="ow">in</span> <span class="n">bands</span><span class="p">])</span>
-
- <span class="c1"># get labels and corresponding colors</span>
- <span class="n">ulabels</span> <span class="o">=</span> <span class="p">[</span><span class="n">label</span><span class="p">[</span><span class="s1">'label'</span><span class="p">]</span> <span class="k">for</span> <span class="n">label</span> <span class="ow">in</span> <span class="n">labels</span><span class="o">.</span><span class="n">values</span><span class="p">()]</span>
- <span class="n">colors</span> <span class="o">=</span> <span class="p">[</span><span class="n">label</span><span class="p">[</span><span class="s1">'color'</span><span class="p">]</span> <span class="k">for</span> <span class="n">label</span> <span class="ow">in</span> <span class="n">labels</span><span class="o">.</span><span class="n">values</span><span class="p">()]</span>
-
- <span class="c1"># create a ListedColormap</span>
- <span class="n">cmap</span> <span class="o">=</span> <span class="n">ListedColormap</span><span class="p">(</span><span class="n">colors</span><span class="p">)</span>
- <span class="n">boundaries</span> <span class="o">=</span> <span class="p">[</span><span class="o">*</span><span class="n">labels</span><span class="o">.</span><span class="n">keys</span><span class="p">(),</span> <span class="n">cmap</span><span class="o">.</span><span class="n">N</span><span class="p">]</span>
- <span class="n">norm</span> <span class="o">=</span> <span class="n">BoundaryNorm</span><span class="p">(</span><span class="n">boundaries</span><span class="p">,</span> <span class="n">cmap</span><span class="o">.</span><span class="n">N</span><span class="p">)</span>
-
- <span class="c1"># create a patch (proxy artist) for every color</span>
- <span class="n">patches</span> <span class="o">=</span> <span class="p">[</span><span class="n">mpatches</span><span class="o">.</span><span class="n">Patch</span><span class="p">(</span><span class="n">color</span><span class="o">=</span><span class="n">c</span><span class="p">,</span> <span class="n">label</span><span class="o">=</span><span class="n">l</span><span class="p">)</span> <span class="k">for</span> <span class="n">c</span><span class="p">,</span> <span class="n">l</span> <span class="ow">in</span>
- <span class="nb">zip</span><span class="p">(</span><span class="n">colors</span><span class="p">,</span> <span class="n">ulabels</span><span class="p">)]</span>
-
- <span class="c1"># initialize figure</span>
- <span class="n">fig</span><span class="p">,</span> <span class="n">ax</span> <span class="o">=</span> <span class="n">plt</span><span class="o">.</span><span class="n">subplots</span><span class="p">(</span><span class="mi">1</span><span class="p">,</span> <span class="mi">3</span><span class="p">,</span> <span class="n">figsize</span><span class="o">=</span><span class="n">figsize</span><span class="p">)</span>
-
- <span class="c1"># plot false color composite</span>
- <span class="n">ax</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span><span class="o">.</span><span class="n">imshow</span><span class="p">(</span><span class="n">rgb</span><span class="p">)</span>
- <span class="n">ax</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span><span class="o">.</span><span class="n">set_title</span><span class="p">(</span><span class="s1">'R = </span><span class="si">{}</span><span class="s1">, G = </span><span class="si">{}</span><span class="s1">, B = </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="o">*</span><span class="n">bands</span><span class="p">),</span> <span class="n">pad</span><span class="o">=</span><span class="mi">15</span><span class="p">)</span>
-
- <span class="c1"># check whether to plot ground truth</span>
- <span class="n">acc</span> <span class="o">=</span> <span class="kc">None</span>
- <span class="k">if</span> <span class="n">y</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
- <span class="c1"># remove axis to plot ground truth from figure</span>
- <span class="n">fig</span><span class="o">.</span><span class="n">delaxes</span><span class="p">(</span><span class="n">ax</span><span class="p">[</span><span class="mi">1</span><span class="p">])</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="c1"># plot ground thruth mask</span>
- <span class="n">ax</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span><span class="o">.</span><span class="n">imshow</span><span class="p">(</span><span class="n">y</span><span class="p">,</span> <span class="n">cmap</span><span class="o">=</span><span class="n">cmap</span><span class="p">,</span> <span class="n">interpolation</span><span class="o">=</span><span class="s1">'nearest'</span><span class="p">,</span> <span class="n">norm</span><span class="o">=</span><span class="n">norm</span><span class="p">)</span>
- <span class="n">ax</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span><span class="o">.</span><span class="n">set_title</span><span class="p">(</span><span class="s1">'Ground truth'</span><span class="p">,</span> <span class="n">pad</span><span class="o">=</span><span class="mi">15</span><span class="p">)</span>
-
- <span class="c1"># check whether to plot model prediction</span>
- <span class="k">if</span> <span class="n">y_pred</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
- <span class="c1"># remove axis to plot model prediction from figure</span>
- <span class="n">fig</span><span class="o">.</span><span class="n">delaxes</span><span class="p">(</span><span class="n">ax</span><span class="p">[</span><span class="mi">2</span><span class="p">])</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="c1"># plot model prediction</span>
- <span class="n">ax</span><span class="p">[</span><span class="mi">2</span><span class="p">]</span><span class="o">.</span><span class="n">imshow</span><span class="p">(</span><span class="n">y_pred</span><span class="p">,</span> <span class="n">cmap</span><span class="o">=</span><span class="n">cmap</span><span class="p">,</span> <span class="n">interpolation</span><span class="o">=</span><span class="s1">'nearest'</span><span class="p">,</span> <span class="n">norm</span><span class="o">=</span><span class="n">norm</span><span class="p">)</span>
-
- <span class="c1"># set title</span>
- <span class="n">title</span> <span class="o">=</span> <span class="s1">'Prediction'</span>
- <span class="k">if</span> <span class="n">y</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
- <span class="n">acc</span> <span class="o">=</span> <span class="n">accuracy_function</span><span class="p">(</span><span class="n">y_pred</span><span class="p">,</span> <span class="n">y</span><span class="p">)</span>
- <span class="n">title</span> <span class="o">+=</span> <span class="s1">' (</span><span class="si">{:.2f}</span><span class="s1">%)'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">acc</span> <span class="o">*</span> <span class="mi">100</span><span class="p">)</span>
- <span class="n">ax</span><span class="p">[</span><span class="mi">2</span><span class="p">]</span><span class="o">.</span><span class="n">set_title</span><span class="p">(</span><span class="n">title</span><span class="p">,</span> <span class="n">pad</span><span class="o">=</span><span class="mi">15</span><span class="p">)</span>
-
- <span class="c1"># if a ground truth or a model prediction is plotted, add legend</span>
- <span class="k">if</span> <span class="nb">len</span><span class="p">(</span><span class="n">fig</span><span class="o">.</span><span class="n">axes</span><span class="p">)</span> <span class="o">></span> <span class="mi">1</span><span class="p">:</span>
- <span class="n">plt</span><span class="o">.</span><span class="n">legend</span><span class="p">(</span><span class="n">handles</span><span class="o">=</span><span class="n">patches</span><span class="p">,</span> <span class="n">bbox_to_anchor</span><span class="o">=</span><span class="p">(</span><span class="mf">1.05</span><span class="p">,</span> <span class="mi">1</span><span class="p">),</span> <span class="n">loc</span><span class="o">=</span><span class="mi">2</span><span class="p">,</span>
- <span class="n">frameon</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
-
- <span class="c1"># save figure</span>
- <span class="k">if</span> <span class="n">state</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
- <span class="n">os</span><span class="o">.</span><span class="n">makedirs</span><span class="p">(</span><span class="n">outpath</span><span class="p">,</span> <span class="n">exist_ok</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
- <span class="n">fig</span><span class="o">.</span><span class="n">savefig</span><span class="p">(</span><span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">outpath</span><span class="p">,</span> <span class="n">state</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="s1">'.pt'</span><span class="p">,</span> <span class="s1">'.png'</span><span class="p">)),</span>
- <span class="n">dpi</span><span class="o">=</span><span class="mi">300</span><span class="p">,</span> <span class="n">bbox_inches</span><span class="o">=</span><span class="s1">'tight'</span><span class="p">)</span>
-
- <span class="k">return</span> <span class="n">fig</span><span class="p">,</span> <span class="n">ax</span></div>
-
-
-<div class="viewcode-block" id="plot_confusion_matrix"><a class="viewcode-back" href="../../source/core.html#core.graphics.plot_confusion_matrix">[docs]</a><span class="k">def</span> <span class="nf">plot_confusion_matrix</span><span class="p">(</span><span class="n">cm</span><span class="p">,</span> <span class="n">labels</span><span class="p">,</span> <span class="n">normalize</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span>
- <span class="n">figsize</span><span class="o">=</span><span class="p">(</span><span class="mi">10</span><span class="p">,</span> <span class="mi">10</span><span class="p">),</span> <span class="n">cmap</span><span class="o">=</span><span class="s1">'Blues'</span><span class="p">,</span> <span class="n">state</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span>
- <span class="n">outpath</span><span class="o">=</span><span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">HERE</span><span class="p">,</span> <span class="s1">'_graphics/'</span><span class="p">)):</span>
- <span class="sd">"""Plot the confusion matrix ``cm``.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> cm : `numpy.ndarray`</span>
-<span class="sd"> The confusion matrix.</span>
-<span class="sd"> labels : `dict` [`int`, `dict`]</span>
-<span class="sd"> The label dictionary. The keys are the values of the class labels</span>
-<span class="sd"> in the ground truth ``y``. Each nested `dict` should have keys:</span>
-<span class="sd"> ``'color'``</span>
-<span class="sd"> A named color (`str`).</span>
-<span class="sd"> ``'label'``</span>
-<span class="sd"> The name of the class label (`str`).</span>
-<span class="sd"> normalize : `bool`, optional</span>
-<span class="sd"> Whether to normalize the confusion matrix. The default is True.</span>
-<span class="sd"> figsize : `tuple`, optional</span>
-<span class="sd"> The figure size in centimeters. The default is (10, 10).</span>
-<span class="sd"> cmap : `str`, optional</span>
-<span class="sd"> A colormap in `matplotlib.pyplot.colormaps()`. The default is 'Blues'.</span>
-<span class="sd"> state : `str` or `None`, optional</span>
-<span class="sd"> Filename to save the plot to. ``state`` should be an existing model</span>
-<span class="sd"> state file ending with '.pt'. The default is None, i.e. plot is not</span>
-<span class="sd"> saved to disk.</span>
-<span class="sd"> outpath : `str` or `pathlib.Path`, optional</span>
-<span class="sd"> Output path. The default is 'pysegcnn/main/_graphics/'.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> fig : `matplotlib.figure.Figure`</span>
-<span class="sd"> The figure handle.</span>
-<span class="sd"> ax : `matplotlib.axes._subplots.AxesSubplot`</span>
-<span class="sd"> The axes handle.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># number of classes</span>
- <span class="n">labels</span> <span class="o">=</span> <span class="p">[</span><span class="n">label</span><span class="p">[</span><span class="s1">'label'</span><span class="p">]</span> <span class="k">for</span> <span class="n">label</span> <span class="ow">in</span> <span class="n">labels</span><span class="o">.</span><span class="n">values</span><span class="p">()]</span>
- <span class="n">nclasses</span> <span class="o">=</span> <span class="nb">len</span><span class="p">(</span><span class="n">labels</span><span class="p">)</span>
-
- <span class="c1"># string format to plot values of confusion matrix</span>
- <span class="n">fmt</span> <span class="o">=</span> <span class="s1">'.0f'</span>
-
- <span class="c1"># minimum and maximum values of the colorbar</span>
- <span class="n">vmin</span><span class="p">,</span> <span class="n">vmax</span> <span class="o">=</span> <span class="mi">0</span><span class="p">,</span> <span class="n">cm</span><span class="o">.</span><span class="n">max</span><span class="p">()</span>
-
- <span class="c1"># check whether to normalize the confusion matrix</span>
- <span class="k">if</span> <span class="n">normalize</span><span class="p">:</span>
- <span class="c1"># normalize</span>
- <span class="n">norm</span> <span class="o">=</span> <span class="n">cm</span><span class="o">.</span><span class="n">sum</span><span class="p">(</span><span class="n">axis</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">keepdims</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
-
- <span class="c1"># check for division by zero</span>
- <span class="n">norm</span><span class="p">[</span><span class="n">norm</span> <span class="o">==</span> <span class="mi">0</span><span class="p">]</span> <span class="o">=</span> <span class="mi">1</span>
- <span class="n">cm</span> <span class="o">=</span> <span class="n">cm</span> <span class="o">/</span> <span class="n">norm</span>
-
- <span class="c1"># change string format to floating point</span>
- <span class="n">fmt</span> <span class="o">=</span> <span class="s1">'.2f'</span>
- <span class="n">vmin</span><span class="p">,</span> <span class="n">vmax</span> <span class="o">=</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">1</span>
-
- <span class="c1"># create figure</span>
- <span class="n">fig</span><span class="p">,</span> <span class="n">ax</span> <span class="o">=</span> <span class="n">plt</span><span class="o">.</span><span class="n">subplots</span><span class="p">(</span><span class="mi">1</span><span class="p">,</span> <span class="mi">1</span><span class="p">,</span> <span class="n">figsize</span><span class="o">=</span><span class="n">figsize</span><span class="p">)</span>
-
- <span class="c1"># get colormap</span>
- <span class="n">cmap</span> <span class="o">=</span> <span class="n">colormap</span><span class="o">.</span><span class="n">get_cmap</span><span class="p">(</span><span class="n">cmap</span><span class="p">,</span> <span class="mi">256</span><span class="p">)</span>
-
- <span class="c1"># plot confusion matrix</span>
- <span class="n">im</span> <span class="o">=</span> <span class="n">ax</span><span class="o">.</span><span class="n">imshow</span><span class="p">(</span><span class="n">cm</span><span class="p">,</span> <span class="n">cmap</span><span class="o">=</span><span class="n">cmap</span><span class="p">,</span> <span class="n">vmin</span><span class="o">=</span><span class="n">vmin</span><span class="p">,</span> <span class="n">vmax</span><span class="o">=</span><span class="n">vmax</span><span class="p">)</span>
-
- <span class="c1"># threshold determining the color of the values</span>
- <span class="n">thresh</span> <span class="o">=</span> <span class="p">(</span><span class="n">cm</span><span class="o">.</span><span class="n">max</span><span class="p">()</span> <span class="o">+</span> <span class="n">cm</span><span class="o">.</span><span class="n">min</span><span class="p">())</span> <span class="o">/</span> <span class="mi">2</span>
-
- <span class="c1"># brightest/darkest color of current colormap</span>
- <span class="n">cmap_min</span><span class="p">,</span> <span class="n">cmap_max</span> <span class="o">=</span> <span class="n">cmap</span><span class="p">(</span><span class="mi">0</span><span class="p">),</span> <span class="n">cmap</span><span class="p">(</span><span class="mi">256</span><span class="p">)</span>
-
- <span class="c1"># plot values of confusion matrix</span>
- <span class="k">for</span> <span class="n">i</span><span class="p">,</span> <span class="n">j</span> <span class="ow">in</span> <span class="n">itertools</span><span class="o">.</span><span class="n">product</span><span class="p">(</span><span class="nb">range</span><span class="p">(</span><span class="n">nclasses</span><span class="p">),</span> <span class="nb">range</span><span class="p">(</span><span class="n">nclasses</span><span class="p">)):</span>
- <span class="n">ax</span><span class="o">.</span><span class="n">text</span><span class="p">(</span><span class="n">j</span><span class="p">,</span> <span class="n">i</span><span class="p">,</span> <span class="nb">format</span><span class="p">(</span><span class="n">cm</span><span class="p">[</span><span class="n">i</span><span class="p">,</span> <span class="n">j</span><span class="p">],</span> <span class="n">fmt</span><span class="p">),</span> <span class="n">ha</span><span class="o">=</span><span class="s1">'center'</span><span class="p">,</span> <span class="n">va</span><span class="o">=</span><span class="s1">'center'</span><span class="p">,</span>
- <span class="n">color</span><span class="o">=</span><span class="n">cmap_max</span> <span class="k">if</span> <span class="n">cm</span><span class="p">[</span><span class="n">i</span><span class="p">,</span> <span class="n">j</span><span class="p">]</span> <span class="o"><</span> <span class="n">thresh</span> <span class="k">else</span> <span class="n">cmap_min</span><span class="p">)</span>
-
- <span class="c1"># axes properties and labels</span>
- <span class="n">ax</span><span class="o">.</span><span class="n">set</span><span class="p">(</span><span class="n">xticks</span><span class="o">=</span><span class="n">np</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="n">nclasses</span><span class="p">),</span>
- <span class="n">yticks</span><span class="o">=</span><span class="n">np</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="n">nclasses</span><span class="p">),</span>
- <span class="n">xticklabels</span><span class="o">=</span><span class="n">labels</span><span class="p">,</span>
- <span class="n">yticklabels</span><span class="o">=</span><span class="n">labels</span><span class="p">,</span>
- <span class="n">ylabel</span><span class="o">=</span><span class="s1">'True'</span><span class="p">,</span>
- <span class="n">xlabel</span><span class="o">=</span><span class="s1">'Predicted'</span><span class="p">)</span>
-
- <span class="c1"># add colorbar axes</span>
- <span class="n">cax</span> <span class="o">=</span> <span class="n">fig</span><span class="o">.</span><span class="n">add_axes</span><span class="p">([</span><span class="n">ax</span><span class="o">.</span><span class="n">get_position</span><span class="p">()</span><span class="o">.</span><span class="n">x1</span> <span class="o">+</span> <span class="mf">0.025</span><span class="p">,</span> <span class="n">ax</span><span class="o">.</span><span class="n">get_position</span><span class="p">()</span><span class="o">.</span><span class="n">y0</span><span class="p">,</span>
- <span class="mf">0.05</span><span class="p">,</span> <span class="n">ax</span><span class="o">.</span><span class="n">get_position</span><span class="p">()</span><span class="o">.</span><span class="n">y1</span> <span class="o">-</span> <span class="n">ax</span><span class="o">.</span><span class="n">get_position</span><span class="p">()</span><span class="o">.</span><span class="n">y0</span><span class="p">])</span>
- <span class="n">fig</span><span class="o">.</span><span class="n">colorbar</span><span class="p">(</span><span class="n">im</span><span class="p">,</span> <span class="n">cax</span><span class="o">=</span><span class="n">cax</span><span class="p">)</span>
-
- <span class="c1"># save figure</span>
- <span class="k">if</span> <span class="n">state</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
- <span class="n">os</span><span class="o">.</span><span class="n">makedirs</span><span class="p">(</span><span class="n">outpath</span><span class="p">,</span> <span class="n">exist_ok</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
- <span class="n">fig</span><span class="o">.</span><span class="n">savefig</span><span class="p">(</span><span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">outpath</span><span class="p">,</span> <span class="n">state</span><span class="p">),</span> <span class="n">dpi</span><span class="o">=</span><span class="mi">300</span><span class="p">,</span> <span class="n">bbox_inches</span><span class="o">=</span><span class="s1">'tight'</span><span class="p">)</span>
-
- <span class="k">return</span> <span class="n">fig</span><span class="p">,</span> <span class="n">ax</span></div>
-
-
-<div class="viewcode-block" id="plot_loss"><a class="viewcode-back" href="../../source/core.html#core.graphics.plot_loss">[docs]</a><span class="k">def</span> <span class="nf">plot_loss</span><span class="p">(</span><span class="n">state_file</span><span class="p">,</span> <span class="n">figsize</span><span class="o">=</span><span class="p">(</span><span class="mi">10</span><span class="p">,</span> <span class="mi">10</span><span class="p">),</span> <span class="n">step</span><span class="o">=</span><span class="mi">5</span><span class="p">,</span>
- <span class="n">colors</span><span class="o">=</span><span class="p">[</span><span class="s1">'lightgreen'</span><span class="p">,</span> <span class="s1">'green'</span><span class="p">,</span> <span class="s1">'skyblue'</span><span class="p">,</span> <span class="s1">'steelblue'</span><span class="p">],</span>
- <span class="n">outpath</span><span class="o">=</span><span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">HERE</span><span class="p">,</span> <span class="s1">'_graphics/'</span><span class="p">)):</span>
- <span class="sd">"""Plot the observed loss and accuracy of a model run.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> state_file : `str` or `pathlib.Path`</span>
-<span class="sd"> The model state file. Model state files are stored in</span>
-<span class="sd"> pysegcnn/main/_models.</span>
-<span class="sd"> figsize : `tuple`, optional</span>
-<span class="sd"> The figure size in centimeters. The default is (10, 10).</span>
-<span class="sd"> step : `int`, optional</span>
-<span class="sd"> The step of epochs for the x-axis labels. The default is 5, i.e. label</span>
-<span class="sd"> each fifth epoch.</span>
-<span class="sd"> colors : `list` [`str`], optional</span>
-<span class="sd"> A list of four named colors supported by `matplotlib`.</span>
-<span class="sd"> The default is ['lightgreen', 'green', 'skyblue', 'steelblue'].</span>
-<span class="sd"> outpath : `str` or `pathlib.Path`, optional</span>
-<span class="sd"> Output path. The default is 'pysegcnn/main/_graphics/'.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> fig : `matplotlib.figure.Figure`</span>
-<span class="sd"> The figure handle.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># load the model state</span>
- <span class="n">model_state</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">load</span><span class="p">(</span><span class="n">state_file</span><span class="p">)</span>
-
- <span class="c1"># get all non-zero elements, i.e. get number of epochs trained before</span>
- <span class="c1"># early stop</span>
- <span class="n">loss</span> <span class="o">=</span> <span class="p">{</span><span class="n">k</span><span class="p">:</span> <span class="n">v</span><span class="p">[</span><span class="n">np</span><span class="o">.</span><span class="n">nonzero</span><span class="p">(</span><span class="n">v</span><span class="p">)]</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">v</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="o">-</span><span class="mi">1</span><span class="p">)</span> <span class="k">for</span> <span class="n">k</span><span class="p">,</span> <span class="n">v</span> <span class="ow">in</span>
- <span class="n">model_state</span><span class="p">[</span><span class="s1">'state'</span><span class="p">]</span><span class="o">.</span><span class="n">items</span><span class="p">()}</span>
-
- <span class="c1"># compute running mean with a window equal to the number of batches in</span>
- <span class="c1"># an epoch</span>
- <span class="n">rm</span> <span class="o">=</span> <span class="p">{</span><span class="n">k</span><span class="p">:</span> <span class="n">running_mean</span><span class="p">(</span><span class="n">v</span><span class="o">.</span><span class="n">flatten</span><span class="p">(</span><span class="s1">'F'</span><span class="p">),</span> <span class="n">v</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">0</span><span class="p">])</span> <span class="k">for</span> <span class="n">k</span><span class="p">,</span> <span class="n">v</span> <span class="ow">in</span> <span class="n">loss</span><span class="o">.</span><span class="n">items</span><span class="p">()}</span>
-
- <span class="c1"># sort the keys of the dictionary alphabetically</span>
- <span class="n">rm</span> <span class="o">=</span> <span class="p">{</span><span class="n">k</span><span class="p">:</span> <span class="n">rm</span><span class="p">[</span><span class="n">k</span><span class="p">]</span> <span class="k">for</span> <span class="n">k</span> <span class="ow">in</span> <span class="nb">sorted</span><span class="p">(</span><span class="n">rm</span><span class="p">)}</span>
-
- <span class="c1"># number of epochs trained</span>
- <span class="n">epochs</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="n">loss</span><span class="p">[</span><span class="s1">'tl'</span><span class="p">]</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">1</span><span class="p">])</span>
-
- <span class="c1"># instanciate figure</span>
- <span class="n">fig</span><span class="p">,</span> <span class="n">ax1</span> <span class="o">=</span> <span class="n">plt</span><span class="o">.</span><span class="n">subplots</span><span class="p">(</span><span class="mi">1</span><span class="p">,</span> <span class="mi">1</span><span class="p">,</span> <span class="n">figsize</span><span class="o">=</span><span class="n">figsize</span><span class="p">)</span>
-
- <span class="c1"># create axes for each parameter to plot</span>
- <span class="n">ax2</span> <span class="o">=</span> <span class="n">ax1</span><span class="o">.</span><span class="n">twinx</span><span class="p">()</span>
- <span class="n">ax3</span> <span class="o">=</span> <span class="n">ax1</span><span class="o">.</span><span class="n">twiny</span><span class="p">()</span>
- <span class="n">ax4</span> <span class="o">=</span> <span class="n">ax2</span><span class="o">.</span><span class="n">twiny</span><span class="p">()</span>
-
- <span class="c1"># list of axes</span>
- <span class="n">axes</span> <span class="o">=</span> <span class="p">[</span><span class="n">ax2</span><span class="p">,</span> <span class="n">ax1</span><span class="p">,</span> <span class="n">ax4</span><span class="p">,</span> <span class="n">ax3</span><span class="p">]</span>
-
- <span class="c1"># plot running mean loss and accuracy of the training dataset</span>
- <span class="p">[</span><span class="n">ax</span><span class="o">.</span><span class="n">plot</span><span class="p">(</span><span class="n">v</span><span class="p">,</span> <span class="n">color</span><span class="o">=</span><span class="n">c</span><span class="p">)</span> <span class="k">for</span> <span class="p">(</span><span class="n">k</span><span class="p">,</span> <span class="n">v</span><span class="p">),</span> <span class="n">ax</span><span class="p">,</span> <span class="n">c</span> <span class="ow">in</span> <span class="nb">zip</span><span class="p">(</span><span class="n">rm</span><span class="o">.</span><span class="n">items</span><span class="p">(),</span> <span class="n">axes</span><span class="p">,</span> <span class="n">colors</span><span class="p">)</span>
- <span class="k">if</span> <span class="n">v</span><span class="o">.</span><span class="n">any</span><span class="p">()]</span>
-
- <span class="c1"># axes properties and labels</span>
- <span class="n">nbatches</span> <span class="o">=</span> <span class="n">loss</span><span class="p">[</span><span class="s1">'tl'</span><span class="p">]</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span>
- <span class="n">ax3</span><span class="o">.</span><span class="n">set</span><span class="p">(</span><span class="n">xticks</span><span class="o">=</span><span class="p">[],</span> <span class="n">xticklabels</span><span class="o">=</span><span class="p">[])</span>
- <span class="n">ax4</span><span class="o">.</span><span class="n">set</span><span class="p">(</span><span class="n">xticks</span><span class="o">=</span><span class="p">[],</span> <span class="n">xticklabels</span><span class="o">=</span><span class="p">[])</span>
- <span class="n">ax1</span><span class="o">.</span><span class="n">set</span><span class="p">(</span><span class="n">xticks</span><span class="o">=</span><span class="n">np</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="n">nbatches</span> <span class="o">*</span> <span class="n">epochs</span><span class="p">[</span><span class="o">-</span><span class="mi">1</span><span class="p">]</span> <span class="o">+</span> <span class="mi">1</span><span class="p">,</span> <span class="n">nbatches</span> <span class="o">*</span> <span class="n">step</span><span class="p">),</span>
- <span class="n">xticklabels</span><span class="o">=</span><span class="n">epochs</span><span class="p">[::</span><span class="n">step</span><span class="p">],</span>
- <span class="n">xlabel</span><span class="o">=</span><span class="s1">'Epoch'</span><span class="p">,</span>
- <span class="n">ylabel</span><span class="o">=</span><span class="s1">'Loss'</span><span class="p">,</span>
- <span class="n">ylim</span><span class="o">=</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="mi">1</span><span class="p">))</span>
- <span class="n">ax2</span><span class="o">.</span><span class="n">set</span><span class="p">(</span><span class="n">ylabel</span><span class="o">=</span><span class="s1">'Accuracy'</span><span class="p">,</span>
- <span class="n">ylim</span><span class="o">=</span><span class="p">(</span><span class="mf">0.5</span><span class="p">,</span> <span class="mi">1</span><span class="p">))</span>
-
- <span class="c1"># compute early stopping point</span>
- <span class="k">if</span> <span class="n">loss</span><span class="p">[</span><span class="s1">'va'</span><span class="p">]</span><span class="o">.</span><span class="n">any</span><span class="p">():</span>
- <span class="n">esepoch</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">argmax</span><span class="p">(</span><span class="n">loss</span><span class="p">[</span><span class="s1">'va'</span><span class="p">]</span><span class="o">.</span><span class="n">mean</span><span class="p">(</span><span class="n">axis</span><span class="o">=</span><span class="mi">0</span><span class="p">))</span> <span class="o">*</span> <span class="n">nbatches</span> <span class="o">+</span> <span class="mi">1</span>
- <span class="n">esacc</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">max</span><span class="p">(</span><span class="n">loss</span><span class="p">[</span><span class="s1">'va'</span><span class="p">]</span><span class="o">.</span><span class="n">mean</span><span class="p">(</span><span class="n">axis</span><span class="o">=</span><span class="mi">0</span><span class="p">))</span>
- <span class="n">ax1</span><span class="o">.</span><span class="n">vlines</span><span class="p">(</span><span class="n">esepoch</span><span class="p">,</span> <span class="n">ymin</span><span class="o">=</span><span class="n">ax1</span><span class="o">.</span><span class="n">get_ylim</span><span class="p">()[</span><span class="mi">0</span><span class="p">],</span> <span class="n">ymax</span><span class="o">=</span><span class="n">ax1</span><span class="o">.</span><span class="n">get_ylim</span><span class="p">()[</span><span class="mi">1</span><span class="p">],</span>
- <span class="n">ls</span><span class="o">=</span><span class="s1">'--'</span><span class="p">,</span> <span class="n">color</span><span class="o">=</span><span class="s1">'grey'</span><span class="p">)</span>
- <span class="n">ax1</span><span class="o">.</span><span class="n">text</span><span class="p">(</span><span class="n">esepoch</span> <span class="o">-</span> <span class="n">nbatches</span><span class="p">,</span> <span class="n">ax1</span><span class="o">.</span><span class="n">get_ylim</span><span class="p">()[</span><span class="mi">0</span><span class="p">]</span> <span class="o">+</span> <span class="mf">0.01</span><span class="p">,</span>
- <span class="s1">'epoch = </span><span class="si">{}</span><span class="s1">, accuracy = </span><span class="si">{:.1f}</span><span class="s1">%'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="nb">int</span><span class="p">(</span><span class="n">esepoch</span> <span class="o">/</span> <span class="n">nbatches</span><span class="p">),</span> <span class="n">esacc</span> <span class="o">*</span> <span class="mi">100</span><span class="p">),</span>
- <span class="n">ha</span><span class="o">=</span><span class="s1">'right'</span><span class="p">,</span> <span class="n">color</span><span class="o">=</span><span class="s1">'grey'</span><span class="p">)</span>
-
- <span class="c1"># create a patch (proxy artist) for every color</span>
- <span class="n">ulabels</span> <span class="o">=</span> <span class="p">[</span><span class="s1">'Training accuracy'</span><span class="p">,</span> <span class="s1">'Training loss'</span><span class="p">,</span>
- <span class="s1">'Validation accuracy'</span><span class="p">,</span> <span class="s1">'Validation loss'</span><span class="p">]</span>
- <span class="n">patches</span> <span class="o">=</span> <span class="p">[</span><span class="n">mlines</span><span class="o">.</span><span class="n">Line2D</span><span class="p">([],</span> <span class="p">[],</span> <span class="n">color</span><span class="o">=</span><span class="n">c</span><span class="p">,</span> <span class="n">label</span><span class="o">=</span><span class="n">l</span><span class="p">)</span> <span class="k">for</span> <span class="n">c</span><span class="p">,</span> <span class="n">l</span> <span class="ow">in</span>
- <span class="nb">zip</span><span class="p">(</span><span class="n">colors</span><span class="p">,</span> <span class="n">ulabels</span><span class="p">)]</span>
- <span class="c1"># plot patches as legend</span>
- <span class="n">ax1</span><span class="o">.</span><span class="n">legend</span><span class="p">(</span><span class="n">handles</span><span class="o">=</span><span class="n">patches</span><span class="p">,</span> <span class="n">loc</span><span class="o">=</span><span class="s1">'lower left'</span><span class="p">,</span> <span class="n">frameon</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
-
- <span class="c1"># save figure</span>
- <span class="n">os</span><span class="o">.</span><span class="n">makedirs</span><span class="p">(</span><span class="n">outpath</span><span class="p">,</span> <span class="n">exist_ok</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
- <span class="n">fig</span><span class="o">.</span><span class="n">savefig</span><span class="p">(</span><span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">join</span><span class="p">(</span>
- <span class="n">outpath</span><span class="p">,</span> <span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">basename</span><span class="p">(</span><span class="n">state_file</span><span class="p">)</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="s1">'.pt'</span><span class="p">,</span> <span class="s1">'.png'</span><span class="p">)),</span>
- <span class="n">dpi</span><span class="o">=</span><span class="mi">300</span><span class="p">,</span> <span class="n">bbox_inches</span><span class="o">=</span><span class="s1">'tight'</span><span class="p">)</span>
-
- <span class="k">return</span> <span class="n">fig</span></div>
-</pre></div>
-
- </div>
-
- </div>
- <footer>
-
-
- <hr/>
-
- <div role="contentinfo">
- <p>
-
- © Copyright 2020, Daniel Frisinghelli
-
- </p>
- </div>
-
-
-
- Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a
-
- <a href="https://github.com/rtfd/sphinx_rtd_theme">theme</a>
-
- provided by <a href="https://readthedocs.org">Read the Docs</a>.
-
-</footer>
-
- </div>
- </div>
-
- </section>
-
- </div>
-
-
- <script type="text/javascript">
- jQuery(function () {
- SphinxRtdTheme.Navigation.enable(true);
- });
- </script>
-
-
-
-
-
-
-</body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/_modules/core/layers.html b/docs/_build/html/_modules/core/layers.html
deleted file mode 100644
index 81af4ab31d865ed5ebee3be36baef12f8dd5c715..0000000000000000000000000000000000000000
--- a/docs/_build/html/_modules/core/layers.html
+++ /dev/null
@@ -1,954 +0,0 @@
-
-
-<!DOCTYPE html>
-<html class="writer-html5" lang="en" >
-<head>
- <meta charset="utf-8">
-
- <meta name="viewport" content="width=device-width, initial-scale=1.0">
-
- <title>core.layers — PySegCNN documentation</title>
-
-
-
- <link rel="stylesheet" href="../../_static/css/theme.css" type="text/css" />
- <link rel="stylesheet" href="../../_static/pygments.css" type="text/css" />
-
-
-
-
-
-
-
- <!--[if lt IE 9]>
- <script src="../../_static/js/html5shiv.min.js"></script>
- <![endif]-->
-
-
- <script type="text/javascript" id="documentation_options" data-url_root="../../" src="../../_static/documentation_options.js"></script>
- <script src="../../_static/jquery.js"></script>
- <script src="../../_static/underscore.js"></script>
- <script src="../../_static/doctools.js"></script>
- <script src="../../_static/language_data.js"></script>
-
- <script type="text/javascript" src="../../_static/js/theme.js"></script>
-
-
- <link rel="index" title="Index" href="../../genindex.html" />
- <link rel="search" title="Search" href="../../search.html" />
-</head>
-
-<body class="wy-body-for-nav">
-
-
- <div class="wy-grid-for-nav">
-
- <nav data-toggle="wy-nav-shift" class="wy-nav-side">
- <div class="wy-side-scroll">
- <div class="wy-side-nav-search" >
-
-
-
- <a href="../../index.html" class="icon icon-home" alt="Documentation Home"> PySegCNN
-
-
-
- </a>
-
-
-
-
-
-
-
-<div role="search">
- <form id="rtd-search-form" class="wy-form" action="../../search.html" method="get">
- <input type="text" name="q" placeholder="Search docs" />
- <input type="hidden" name="check_keywords" value="yes" />
- <input type="hidden" name="area" value="default" />
- </form>
-</div>
-
-
- </div>
-
-
- <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
-
-
-
-
-
-
- <ul>
-<li class="toctree-l1"><a class="reference internal" href="../../source/requirements.html">Requirements</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../../source/installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../../source/api.html">API Reference</a></li>
-</ul>
-
-
-
- </div>
-
- </div>
- </nav>
-
- <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
-
-
- <nav class="wy-nav-top" aria-label="top navigation">
-
- <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
- <a href="../../index.html">PySegCNN</a>
-
- </nav>
-
-
- <div class="wy-nav-content">
-
- <div class="rst-content">
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-<div role="navigation" aria-label="breadcrumbs navigation">
-
- <ul class="wy-breadcrumbs">
-
- <li><a href="../../index.html" class="icon icon-home"></a> »</li>
-
- <li><a href="../index.html">Module code</a> »</li>
-
- <li>core.layers</li>
-
-
- <li class="wy-breadcrumbs-aside">
-
- </li>
-
- </ul>
-
-
- <hr/>
-</div>
- <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
- <div itemprop="articleBody">
-
- <h1>Source code for core.layers</h1><div class="highlight"><pre>
-<span></span><span class="sd">"""Layers of a convolutional encoder-decoder network.</span>
-
-<span class="sd">License</span>
-<span class="sd">-------</span>
-
-<span class="sd"> Copyright (c) 2020 Daniel Frisinghelli</span>
-
-<span class="sd"> This source code is licensed under the GNU General Public License v3.</span>
-
-<span class="sd"> See the LICENSE file in the repository's root directory.</span>
-
-<span class="sd">"""</span>
-
-<span class="c1"># !/usr/bin/env python</span>
-<span class="c1"># -*- coding: utf-8 -*-</span>
-
-<span class="c1"># externals</span>
-<span class="kn">import</span> <span class="nn">numpy</span> <span class="k">as</span> <span class="nn">np</span>
-<span class="kn">import</span> <span class="nn">torch</span>
-<span class="kn">import</span> <span class="nn">torch.nn</span> <span class="k">as</span> <span class="nn">nn</span>
-<span class="kn">import</span> <span class="nn">torch.nn.functional</span> <span class="k">as</span> <span class="nn">F</span>
-
-
-<div class="viewcode-block" id="Conv2dSame"><a class="viewcode-back" href="../../source/core.html#core.layers.Conv2dSame">[docs]</a><span class="k">class</span> <span class="nc">Conv2dSame</span><span class="p">(</span><span class="n">nn</span><span class="o">.</span><span class="n">Conv2d</span><span class="p">):</span>
- <span class="sd">"""A convolution preserving the shape of its input.</span>
-
-<span class="sd"> Given the kernel size, the dilation and a stride of 1, the padding is</span>
-<span class="sd"> calculated such that the output of the convolution has the same spatial</span>
-<span class="sd"> dimensions as the input.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> *args: `list` [`str`]</span>
-<span class="sd"> positional arguments passed to `torch.nn.Conv2d`:</span>
-<span class="sd"> ``'in_channels'``: `int`</span>
-<span class="sd"> Number of input channels.</span>
-<span class="sd"> ``'out_channels'``: `int`</span>
-<span class="sd"> Number of output channels.</span>
-<span class="sd"> ``'kernel_size'``: `int` or `tuple` [`int`]</span>
-<span class="sd"> Size of the convolving kernel.</span>
-<span class="sd"> **kwargs: 'dict' [`str`]</span>
-<span class="sd"> Additional keyword arguments passed to `torch.nn.Conv2d`_.</span>
-
-<span class="sd"> .. _torch.nn.Conv2d:</span>
-<span class="sd"> https://pytorch.org/docs/stable/generated/torch.nn.Conv2d.html#torch.nn.Conv2d</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="o">*</span><span class="n">args</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="o">*</span><span class="n">args</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span>
-
- <span class="c1"># define tensorflows "SAME" padding for stride = 1</span>
- <span class="n">x_pad</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">same_padding</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">dilation</span><span class="p">[</span><span class="mi">1</span><span class="p">],</span> <span class="bp">self</span><span class="o">.</span><span class="n">kernel_size</span><span class="p">[</span><span class="mi">1</span><span class="p">])</span>
- <span class="n">y_pad</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">same_padding</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">dilation</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="bp">self</span><span class="o">.</span><span class="n">kernel_size</span><span class="p">[</span><span class="mi">0</span><span class="p">])</span>
-
- <span class="bp">self</span><span class="o">.</span><span class="n">padding</span> <span class="o">=</span> <span class="p">(</span><span class="n">y_pad</span><span class="p">,</span> <span class="n">x_pad</span><span class="p">)</span>
-
-<div class="viewcode-block" id="Conv2dSame.same_padding"><a class="viewcode-back" href="../../source/core.html#core.layers.Conv2dSame.same_padding">[docs]</a> <span class="nd">@staticmethod</span>
- <span class="k">def</span> <span class="nf">same_padding</span><span class="p">(</span><span class="n">d</span><span class="p">,</span> <span class="n">k</span><span class="p">):</span>
- <span class="sd">"""Calculate the amount of padding.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> d : `int`</span>
-<span class="sd"> The dilation of the convolution.</span>
-<span class="sd"> k : `int`</span>
-<span class="sd"> The kernel size.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> p : `int`</span>
-<span class="sd"> the amount of padding.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># calculates the padding so that the convolution</span>
- <span class="c1"># conserves the shape of its input when stride = 1</span>
- <span class="k">return</span> <span class="nb">int</span><span class="p">(</span><span class="n">d</span> <span class="o">*</span> <span class="p">(</span><span class="n">k</span> <span class="o">-</span> <span class="mi">1</span><span class="p">)</span> <span class="o">/</span> <span class="mi">2</span><span class="p">)</span></div></div>
-
-
-<div class="viewcode-block" id="conv_bn_relu"><a class="viewcode-back" href="../../source/core.html#core.layers.conv_bn_relu">[docs]</a><span class="k">def</span> <span class="nf">conv_bn_relu</span><span class="p">(</span><span class="n">in_channels</span><span class="p">,</span> <span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
- <span class="sd">"""Block of convolution, batch normalization and rectified linear unit.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> in_channels : `int`</span>
-<span class="sd"> Number of input channels.</span>
-<span class="sd"> out_channels : `int`</span>
-<span class="sd"> Number of output channels.</span>
-<span class="sd"> **kwargs: 'dict' [`str`]</span>
-<span class="sd"> Additional arguments passed to `pysegcnn.core.layers.Conv2dSame`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> block : `torch.nn.Sequential` [`torch.nn.Module`]</span>
-<span class="sd"> An instance of `torch.nn.Sequential` containing a sequence of</span>
-<span class="sd"> convolution, batch normalization and rectified linear unit layers.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="n">nn</span><span class="o">.</span><span class="n">Sequential</span><span class="p">(</span>
- <span class="n">Conv2dSame</span><span class="p">(</span><span class="n">in_channels</span><span class="p">,</span> <span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">),</span>
- <span class="n">nn</span><span class="o">.</span><span class="n">BatchNorm2d</span><span class="p">(</span><span class="n">out_channels</span><span class="p">),</span>
- <span class="n">nn</span><span class="o">.</span><span class="n">ReLU</span><span class="p">(),</span>
- <span class="n">Conv2dSame</span><span class="p">(</span><span class="n">out_channels</span><span class="p">,</span> <span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">),</span>
- <span class="n">nn</span><span class="o">.</span><span class="n">BatchNorm2d</span><span class="p">(</span><span class="n">out_channels</span><span class="p">),</span>
- <span class="n">nn</span><span class="o">.</span><span class="n">ReLU</span><span class="p">()</span>
- <span class="p">)</span></div>
-
-
-<div class="viewcode-block" id="Block"><a class="viewcode-back" href="../../source/core.html#core.layers.Block">[docs]</a><span class="k">class</span> <span class="nc">Block</span><span class="p">(</span><span class="n">nn</span><span class="o">.</span><span class="n">Module</span><span class="p">):</span>
- <span class="sd">"""Basic convolutional block.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> in_channels : `int`</span>
-<span class="sd"> Number of input channels.</span>
-<span class="sd"> out_channels : `int`</span>
-<span class="sd"> Number of output channels.</span>
-<span class="sd"> **kwargs: 'dict' [`str`]</span>
-<span class="sd"> Additional arguments passed to `pysegcnn.core.layers.Conv2dSame`.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> TypeError</span>
-<span class="sd"> Raised if `~pysegcnn.core.layers.Block.layers` method does not return</span>
-<span class="sd"> an instance of `torch.nn.Sequential`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">in_channels</span><span class="p">,</span> <span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">()</span>
-
- <span class="c1"># number of input and output channels</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">in_channels</span> <span class="o">=</span> <span class="n">in_channels</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">out_channels</span> <span class="o">=</span> <span class="n">out_channels</span>
-
- <span class="c1"># keyword arguments configuring convolutions</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">kwargs</span> <span class="o">=</span> <span class="n">kwargs</span>
-
- <span class="c1"># the layers of the block</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">conv</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">layers</span><span class="p">()</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="nb">isinstance</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">conv</span><span class="p">,</span> <span class="n">nn</span><span class="o">.</span><span class="n">Sequential</span><span class="p">):</span>
- <span class="k">raise</span> <span class="ne">TypeError</span><span class="p">(</span><span class="s1">'</span><span class="si">{}</span><span class="s1">.layers() should return an instance of </span><span class="si">{}</span><span class="s1">.'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="vm">__class__</span><span class="o">.</span><span class="vm">__name__</span><span class="p">,</span>
- <span class="nb">repr</span><span class="p">(</span><span class="n">nn</span><span class="o">.</span><span class="n">Sequential</span><span class="p">)))</span>
-
-<div class="viewcode-block" id="Block.layers"><a class="viewcode-back" href="../../source/core.html#core.layers.Block.layers">[docs]</a> <span class="k">def</span> <span class="nf">layers</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Define the layers of the block.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> NotImplementedError</span>
-<span class="sd"> Raised if `pysegcnn.core.layers.Block` is not inherited.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> layers : `torch.nn.Sequential` [`torch.nn.Module`]</span>
-<span class="sd"> Return an instance of `torch.nn.Sequential` containing a sequence</span>
-<span class="sd"> of layer (`torch.nn.Module` ) instances.</span>
-
-<span class="sd"> """</span>
- <span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">(</span><span class="s1">'Return an instance of </span><span class="si">{}</span><span class="s1">.'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="nb">repr</span><span class="p">(</span><span class="n">nn</span><span class="o">.</span><span class="n">Sequential</span><span class="p">)))</span></div>
-
-<div class="viewcode-block" id="Block.forward"><a class="viewcode-back" href="../../source/core.html#core.layers.Block.forward">[docs]</a> <span class="k">def</span> <span class="nf">forward</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Forward pass of the block.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> NotImplementedError</span>
-<span class="sd"> Raised if `pysegcnn.core.layers.Block` is not inherited.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
- <span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">(</span><span class="s1">'Implement the forward pass.'</span><span class="p">)</span></div></div>
-
-
-<div class="viewcode-block" id="EncoderBlock"><a class="viewcode-back" href="../../source/core.html#core.layers.EncoderBlock">[docs]</a><span class="k">class</span> <span class="nc">EncoderBlock</span><span class="p">(</span><span class="n">Block</span><span class="p">):</span>
- <span class="sd">"""Block of a convolutional encoder.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> in_channels : `int`</span>
-<span class="sd"> Number of input channels.</span>
-<span class="sd"> out_channels : `int`</span>
-<span class="sd"> Number of output channels.</span>
-<span class="sd"> **kwargs: 'dict' [`str`]</span>
-<span class="sd"> Additional arguments passed to `pysegcnn.core.layers.Conv2dSame`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">in_channels</span><span class="p">,</span> <span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">in_channels</span><span class="p">,</span> <span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span>
-
-<div class="viewcode-block" id="EncoderBlock.forward"><a class="viewcode-back" href="../../source/core.html#core.layers.EncoderBlock.forward">[docs]</a> <span class="k">def</span> <span class="nf">forward</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">x</span><span class="p">):</span>
- <span class="sd">"""Forward pass of an encoder block.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> x : `torch.Tensor`</span>
-<span class="sd"> Input tensor, e.g. output of the previous block/layer.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> y : `torch.Tensor`, shape=(batch, channel, height, width)</span>
-<span class="sd"> Output of the encoder block.</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
-<span class="sd"> Intermediate output before applying downsampling. Useful to</span>
-<span class="sd"> implement skip connections.</span>
-<span class="sd"> indices : `torch.Tensor` or `None`</span>
-<span class="sd"> Optional indices of the downsampling method, e.g. indices of the</span>
-<span class="sd"> maxima when using `torch.nn.functional.max_pool2d`. Useful for</span>
-<span class="sd"> upsampling later. If no indices are required to upsample, simply</span>
-<span class="sd"> return ``indices`` = `None`.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># the forward pass of the layers of the block</span>
- <span class="n">x</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">conv</span><span class="p">(</span><span class="n">x</span><span class="p">)</span>
-
- <span class="c1"># the downsampling layer</span>
- <span class="n">y</span><span class="p">,</span> <span class="n">indices</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">downsample</span><span class="p">(</span><span class="n">x</span><span class="p">)</span>
-
- <span class="k">return</span> <span class="p">(</span><span class="n">y</span><span class="p">,</span> <span class="n">x</span><span class="p">,</span> <span class="n">indices</span><span class="p">)</span></div>
-
-<div class="viewcode-block" id="EncoderBlock.downsample"><a class="viewcode-back" href="../../source/core.html#core.layers.EncoderBlock.downsample">[docs]</a> <span class="k">def</span> <span class="nf">downsample</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">x</span><span class="p">):</span>
- <span class="sd">"""Define the downsampling method.</span>
-
-<span class="sd"> The `~pysegcnn.core.layers.EncoderBlock.downsample` method should</span>
-<span class="sd"> implement the spatial pooling operation.</span>
-
-<span class="sd"> Use one of the following functions to downsample:</span>
-<span class="sd"> - `torch.nn.functional.max_pool2d`</span>
-<span class="sd"> - `torch.nn.functional.interpolate`</span>
-
-<span class="sd"> See `pysegcnn.core.layers.ConvBnReluMaxPool` for an example</span>
-<span class="sd"> implementation.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
-<span class="sd"> Input tensor, e.g. output of a convolutional block.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> NotImplementedError</span>
-<span class="sd"> Raised if `pysegcnn.core.layers.EncoderBlock` is not inherited.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
-<span class="sd"> The spatially downsampled tensor.</span>
-<span class="sd"> indices : `torch.Tensor` or `None`</span>
-<span class="sd"> Optional indices of the downsampling method, e.g. indices of the</span>
-<span class="sd"> maxima when using `torch.nn.functional.max_pool2d`. Useful for</span>
-<span class="sd"> upsampling later. If no indices are required to upsample, simply</span>
-<span class="sd"> return ``indices`` = `None`.</span>
-
-<span class="sd"> """</span>
- <span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">(</span><span class="s1">'Implement the downsampling function.'</span><span class="p">)</span></div></div>
-
-
-<div class="viewcode-block" id="DecoderBlock"><a class="viewcode-back" href="../../source/core.html#core.layers.DecoderBlock">[docs]</a><span class="k">class</span> <span class="nc">DecoderBlock</span><span class="p">(</span><span class="n">Block</span><span class="p">):</span>
- <span class="sd">"""Block of a convolutional decoder.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> in_channels : `int`</span>
-<span class="sd"> Number of input channels.</span>
-<span class="sd"> out_channels : `int`</span>
-<span class="sd"> Number of output channels.</span>
-<span class="sd"> **kwargs: 'dict' [`str`]</span>
-<span class="sd"> Additional arguments passed to `pysegcnn.core.layers.Conv2dSame`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">in_channels</span><span class="p">,</span> <span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">in_channels</span><span class="p">,</span> <span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span>
-
-<div class="viewcode-block" id="DecoderBlock.forward"><a class="viewcode-back" href="../../source/core.html#core.layers.DecoderBlock.forward">[docs]</a> <span class="k">def</span> <span class="nf">forward</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">x</span><span class="p">,</span> <span class="n">feature</span><span class="p">,</span> <span class="n">indices</span><span class="p">,</span> <span class="n">skip</span><span class="p">):</span>
- <span class="sd">"""Forward pass of a decoder block.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
-<span class="sd"> Input tensor.</span>
-<span class="sd"> feature : `torch.Tensor`, shape=(batch, channel, height, width)</span>
-<span class="sd"> Intermediate output of a layer in the encoder.</span>
-<span class="sd"> If ``skip`` = True, ``feature`` is concatenated (along the channel</span>
-<span class="sd"> axis) to the output of the respective upsampling layer in the</span>
-<span class="sd"> decoder (skip connection).</span>
-<span class="sd"> indices : `torch.Tensor` or `None`</span>
-<span class="sd"> Indices of the encoder downsampling method.</span>
-<span class="sd"> skip : `bool`</span>
-<span class="sd"> Whether to apply the skip connection.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
-<span class="sd"> Output of the decoder block.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># upsample</span>
- <span class="n">x</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">upsample</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">feature</span><span class="p">,</span> <span class="n">indices</span><span class="p">)</span>
-
- <span class="c1"># check whether to apply the skip connection</span>
- <span class="c1"># skip connection: concatenate the output of a layer in the encoder to</span>
- <span class="c1"># the corresponding layer in the decoder (along the channel axis)</span>
- <span class="k">if</span> <span class="n">skip</span><span class="p">:</span>
- <span class="n">x</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">cat</span><span class="p">([</span><span class="n">x</span><span class="p">,</span> <span class="n">feature</span><span class="p">],</span> <span class="n">axis</span><span class="o">=</span><span class="mi">1</span><span class="p">)</span>
-
- <span class="c1"># output of the convolutional layer</span>
- <span class="n">x</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">conv</span><span class="p">(</span><span class="n">x</span><span class="p">)</span>
-
- <span class="k">return</span> <span class="n">x</span></div>
-
-<div class="viewcode-block" id="DecoderBlock.upsample"><a class="viewcode-back" href="../../source/core.html#core.layers.DecoderBlock.upsample">[docs]</a> <span class="k">def</span> <span class="nf">upsample</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">x</span><span class="p">,</span> <span class="n">feature</span><span class="p">,</span> <span class="n">indices</span><span class="p">):</span>
- <span class="sd">"""Define the upsampling method.</span>
-
-<span class="sd"> The `~pysegcnn.core.layers.DecoderBlock.upsample` method should</span>
-<span class="sd"> implement the spatial upsampling operation.</span>
-
-<span class="sd"> Use one of the following functions to upsample:</span>
-<span class="sd"> - `torch.nn.functional.max_unpool2d`</span>
-<span class="sd"> - `torch.nn.functional.interpolate`</span>
-
-<span class="sd"> See `pysegcnn.core.layers.ConvBnReluMaxUnpool` or</span>
-<span class="sd"> `pysegcnn.core.layers.ConvBnReluUpsample` for an example</span>
-<span class="sd"> implementation.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
-<span class="sd"> Input tensor, e.g. output of a convolutional block.</span>
-<span class="sd"> feature : `torch.Tensor`, shape=(batch, channel, height, width)</span>
-<span class="sd"> Intermediate output of a layer in the encoder. Used to implement</span>
-<span class="sd"> skip connections.</span>
-<span class="sd"> indices : `torch.Tensor` or `None`</span>
-<span class="sd"> Indices of the encoder downsampling method.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> NotImplementedError</span>
-<span class="sd"> Raised if `pysegcnn.core.layers.DecoderBlock` is not inherited.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
-<span class="sd"> The spatially upsampled tensor.</span>
-
-<span class="sd"> """</span>
- <span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">(</span><span class="s1">'Implement the upsampling function'</span><span class="p">)</span></div></div>
-
-
-<div class="viewcode-block" id="Encoder"><a class="viewcode-back" href="../../source/core.html#core.layers.Encoder">[docs]</a><span class="k">class</span> <span class="nc">Encoder</span><span class="p">(</span><span class="n">nn</span><span class="o">.</span><span class="n">Module</span><span class="p">):</span>
- <span class="sd">"""Generic convolutional encoder.</span>
-
-<span class="sd"> When instanciating an encoder-decoder architechure, ``filters`` should be</span>
-<span class="sd"> the same for `pysegcnn.core.layers.Encoder` and</span>
-<span class="sd"> `pysegcnn.core.layers.Decoder`.</span>
-
-<span class="sd"> See `pysegcnn.core.models.UNet` for an example implementation.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> filters : `list` [`int`]</span>
-<span class="sd"> List of input channels to each convolutional block. The length of</span>
-<span class="sd"> ``filters`` determines the depth of the encoder. The first element of</span>
-<span class="sd"> ``filters`` has to be the number of channels of the input images.</span>
-<span class="sd"> block : `pysegcnn.core.layers.EncoderBlock`</span>
-<span class="sd"> The convolutional block defining a layer in the encoder.</span>
-<span class="sd"> A subclass of `pysegcnn.core.layers.EncoderBlock`, e.g.</span>
-<span class="sd"> `pysegcnn.core.layers.ConvBnReluMaxPool`.</span>
-<span class="sd"> **kwargs: 'dict' [`str`]</span>
-<span class="sd"> Additional arguments passed to `pysegcnn.core.layers.Conv2dSame`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">filters</span><span class="p">,</span> <span class="n">block</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">()</span>
-
- <span class="c1"># the number of filters for each block: the first element of filters</span>
- <span class="c1"># has to be the number of input channels</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">features</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">filters</span><span class="p">)</span>
-
- <span class="c1"># the block of operations defining a layer in the encoder</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="nb">issubclass</span><span class="p">(</span><span class="n">block</span><span class="p">,</span> <span class="n">EncoderBlock</span><span class="p">):</span>
- <span class="k">raise</span> <span class="ne">TypeError</span><span class="p">(</span><span class="s1">'"block" expected to be a subclass of </span><span class="si">{}</span><span class="s1">.'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="nb">repr</span><span class="p">(</span><span class="n">EncoderBlock</span><span class="p">)))</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">block</span> <span class="o">=</span> <span class="n">block</span>
-
- <span class="c1"># construct the encoder layers</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">layers</span> <span class="o">=</span> <span class="p">[]</span>
- <span class="k">for</span> <span class="n">lyr</span><span class="p">,</span> <span class="n">lyrp1</span> <span class="ow">in</span> <span class="nb">zip</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">features</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">features</span><span class="p">[</span><span class="mi">1</span><span class="p">:]):</span>
- <span class="c1"># append blocks to the encoder layers</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">layers</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">block</span><span class="p">(</span><span class="n">lyr</span><span class="p">,</span> <span class="n">lyrp1</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">))</span>
-
- <span class="c1"># convert list of layers to ModuleList</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">layers</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">ModuleList</span><span class="p">(</span><span class="o">*</span><span class="p">[</span><span class="bp">self</span><span class="o">.</span><span class="n">layers</span><span class="p">])</span>
-
-<div class="viewcode-block" id="Encoder.forward"><a class="viewcode-back" href="../../source/core.html#core.layers.Encoder.forward">[docs]</a> <span class="k">def</span> <span class="nf">forward</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">x</span><span class="p">):</span>
- <span class="sd">"""Forward pass of the encoder.</span>
-
-<span class="sd"> Stores intermediate outputs in a dictionary. The keys of the dictionary</span>
-<span class="sd"> are the number of the network layers and the values are dictionaries</span>
-<span class="sd"> with the following (key, value) pairs:</span>
-<span class="sd"> ``"feature"``</span>
-<span class="sd"> The intermediate encoder outputs (`torch.Tensor`).</span>
-<span class="sd"> ``"indices"``</span>
-<span class="sd"> The indices of the max pooling layer, if required</span>
-<span class="sd"> (`torch.Tensor`).</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
-<span class="sd"> Input image.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
-<span class="sd"> Output of the encoder.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># initialize a dictionary that caches the intermediate outputs, i.e.</span>
- <span class="c1"># features and pooling indices of each block in the encoder</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">cache</span> <span class="o">=</span> <span class="p">{}</span>
- <span class="k">for</span> <span class="n">i</span><span class="p">,</span> <span class="n">layer</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">layers</span><span class="p">):</span>
- <span class="c1"># apply current encoder layer forward pass</span>
- <span class="n">x</span><span class="p">,</span> <span class="n">y</span><span class="p">,</span> <span class="n">ind</span> <span class="o">=</span> <span class="n">layer</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">x</span><span class="p">)</span>
-
- <span class="c1"># store intermediate outputs for optional skip connections</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">cache</span><span class="p">[</span><span class="n">i</span><span class="p">]</span> <span class="o">=</span> <span class="p">{</span><span class="s1">'feature'</span><span class="p">:</span> <span class="n">y</span><span class="p">,</span> <span class="s1">'indices'</span><span class="p">:</span> <span class="n">ind</span><span class="p">}</span>
-
- <span class="k">return</span> <span class="n">x</span></div></div>
-
-
-<div class="viewcode-block" id="Decoder"><a class="viewcode-back" href="../../source/core.html#core.layers.Decoder">[docs]</a><span class="k">class</span> <span class="nc">Decoder</span><span class="p">(</span><span class="n">nn</span><span class="o">.</span><span class="n">Module</span><span class="p">):</span>
- <span class="sd">"""Generic convolutional decoder.</span>
-
-<span class="sd"> When instanciating an encoder-decoder architechure, ``filters`` should be</span>
-<span class="sd"> the same for `pysegcnn.core.layers.Encoder` and</span>
-<span class="sd"> `pysegcnn.core.layers.Decoder`.</span>
-
-<span class="sd"> See `pysegcnn.core.models.UNet` for an example implementation.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> filters : `list` [`int`]</span>
-<span class="sd"> List of input channels to each convolutional block. The length of</span>
-<span class="sd"> ``filters`` determines the depth of the decoder. The first element of</span>
-<span class="sd"> ``filters`` has to be the number of channels of the input images.</span>
-<span class="sd"> block : `pysegcnn.core.layers.DecoderBlock`</span>
-<span class="sd"> The convolutional block defining a layer in the decoder.</span>
-<span class="sd"> A subclass of `pysegcnn.core.layers.DecoderBlock`, e.g.</span>
-<span class="sd"> `pysegcnn.core.layers.ConvBnReluMaxUnpool`.</span>
-<span class="sd"> skip : `bool`</span>
-<span class="sd"> Whether to apply skip connections from the encoder to the decoder.</span>
-<span class="sd"> **kwargs: 'dict' [`str`]</span>
-<span class="sd"> Additional arguments passed to `pysegcnn.core.layers.Conv2dSame`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">filters</span><span class="p">,</span> <span class="n">block</span><span class="p">,</span> <span class="n">skip</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">()</span>
-
- <span class="c1"># the block of operations defining a layer in the decoder</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="nb">issubclass</span><span class="p">(</span><span class="n">block</span><span class="p">,</span> <span class="n">DecoderBlock</span><span class="p">):</span>
- <span class="k">raise</span> <span class="ne">TypeError</span><span class="p">(</span><span class="s1">'"block" expected to be a subclass of </span><span class="si">{}</span><span class="s1">.'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="nb">repr</span><span class="p">(</span><span class="n">DecoderBlock</span><span class="p">)))</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">block</span> <span class="o">=</span> <span class="n">block</span>
-
- <span class="c1"># the number of filters for each block is symmetric to the encoder:</span>
- <span class="c1"># the last two element of filters have to be equal in order to apply</span>
- <span class="c1"># last skip connection</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">features</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">filters</span><span class="p">)[::</span><span class="o">-</span><span class="mi">1</span><span class="p">]</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">features</span><span class="p">[</span><span class="o">-</span><span class="mi">1</span><span class="p">]</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">features</span><span class="p">[</span><span class="o">-</span><span class="mi">2</span><span class="p">]</span>
-
- <span class="c1"># whether to apply skip connections</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">skip</span> <span class="o">=</span> <span class="n">skip</span>
-
- <span class="c1"># in case of skip connections, the number of input channels to</span>
- <span class="c1"># each block of the decoder is doubled</span>
- <span class="n">n_in</span> <span class="o">=</span> <span class="mi">2</span> <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">skip</span> <span class="k">else</span> <span class="mi">1</span>
-
- <span class="c1"># construct decoder layers</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">layers</span> <span class="o">=</span> <span class="p">[]</span>
- <span class="k">for</span> <span class="n">lyr</span><span class="p">,</span> <span class="n">lyrp1</span> <span class="ow">in</span> <span class="nb">zip</span><span class="p">(</span><span class="n">n_in</span> <span class="o">*</span> <span class="bp">self</span><span class="o">.</span><span class="n">features</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">features</span><span class="p">[</span><span class="mi">1</span><span class="p">:]):</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">layers</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">block</span><span class="p">(</span><span class="n">lyr</span><span class="p">,</span> <span class="n">lyrp1</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">))</span>
-
- <span class="c1"># convert list of layers to ModuleList</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">layers</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">ModuleList</span><span class="p">(</span><span class="o">*</span><span class="p">[</span><span class="bp">self</span><span class="o">.</span><span class="n">layers</span><span class="p">])</span>
-
-<div class="viewcode-block" id="Decoder.forward"><a class="viewcode-back" href="../../source/core.html#core.layers.Decoder.forward">[docs]</a> <span class="k">def</span> <span class="nf">forward</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">x</span><span class="p">,</span> <span class="n">enc_cache</span><span class="p">):</span>
- <span class="sd">"""Forward pass of the decoder.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
-<span class="sd"> Output of the encoder.</span>
-<span class="sd"> enc_cache : `dict` [`dict`]</span>
-<span class="sd"> Cache dictionary. The keys of the dictionary are the number of the</span>
-<span class="sd"> network layers and the values are dictionaries with the following</span>
-<span class="sd"> (key, value) pairs:</span>
-<span class="sd"> ``"feature"``</span>
-<span class="sd"> The intermediate encoder outputs (`torch.Tensor`).</span>
-<span class="sd"> ``"indices"``</span>
-<span class="sd"> The indices of the max pooling layer (`torch.Tensor`).</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
-<span class="sd"> Output of the decoder.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># for each layer, upsample input and apply optional skip connection</span>
- <span class="k">for</span> <span class="n">i</span><span class="p">,</span> <span class="n">layer</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">layers</span><span class="p">):</span>
-
- <span class="c1"># get intermediate outputs from encoder: iterate the encoder cache</span>
- <span class="c1"># in reversed direction, i.e. from last to first encoder layer</span>
- <span class="n">cache</span> <span class="o">=</span> <span class="n">enc_cache</span><span class="p">[</span><span class="nb">len</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">layers</span><span class="p">)</span> <span class="o">-</span> <span class="p">(</span><span class="n">i</span><span class="o">+</span><span class="mi">1</span><span class="p">)]</span>
- <span class="n">feature</span><span class="p">,</span> <span class="n">indices</span> <span class="o">=</span> <span class="n">cache</span><span class="p">[</span><span class="s1">'feature'</span><span class="p">],</span> <span class="n">cache</span><span class="p">[</span><span class="s1">'indices'</span><span class="p">]</span>
-
- <span class="c1"># apply current decoder layer forward pass</span>
- <span class="n">x</span> <span class="o">=</span> <span class="n">layer</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">feature</span><span class="p">,</span> <span class="n">indices</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">skip</span><span class="p">)</span>
-
- <span class="k">return</span> <span class="n">x</span></div></div>
-
-
-<div class="viewcode-block" id="ConvBnReluMaxPool"><a class="viewcode-back" href="../../source/core.html#core.layers.ConvBnReluMaxPool">[docs]</a><span class="k">class</span> <span class="nc">ConvBnReluMaxPool</span><span class="p">(</span><span class="n">EncoderBlock</span><span class="p">):</span>
- <span class="sd">"""Block of convolution, batchnorm, relu and 2x2 max pool.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> in_channels : `int`</span>
-<span class="sd"> Number of input channels.</span>
-<span class="sd"> out_channels : `int`</span>
-<span class="sd"> Number of output channels.</span>
-<span class="sd"> **kwargs: 'dict' [`str`]</span>
-<span class="sd"> Additional keyword arguments passed to</span>
-<span class="sd"> `pysegcnn.core.layers.Conv2dSame`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">in_channels</span><span class="p">,</span> <span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">in_channels</span><span class="p">,</span> <span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span>
-
-<div class="viewcode-block" id="ConvBnReluMaxPool.layers"><a class="viewcode-back" href="../../source/core.html#core.layers.ConvBnReluMaxPool.layers">[docs]</a> <span class="k">def</span> <span class="nf">layers</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Sequence of convolution, batchnorm and relu layers.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> layers : `torch.nn.Sequential` [`torch.nn.Module`]</span>
-<span class="sd"> An instance of `torch.nn.Sequential` containing the sequence</span>
-<span class="sd"> of convolution, batchnorm and relu layer (`torch.nn.Module`)</span>
-<span class="sd"> instances.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="n">conv_bn_relu</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">in_channels</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="bp">self</span><span class="o">.</span><span class="n">kwargs</span><span class="p">)</span></div>
-
-<div class="viewcode-block" id="ConvBnReluMaxPool.downsample"><a class="viewcode-back" href="../../source/core.html#core.layers.ConvBnReluMaxPool.downsample">[docs]</a> <span class="k">def</span> <span class="nf">downsample</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">x</span><span class="p">):</span>
- <span class="sd">"""2x2 max pooling layer, `torch.nn.functional.max_pool2d`.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
-<span class="sd"> Input tensor.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height // 2, width // 2)</span>
-<span class="sd"> The 2x2 max pooled tensor.</span>
-<span class="sd"> indices : `torch.Tensor` or `None`</span>
-<span class="sd"> The indices of the maxima. Useful for upsampling with</span>
-<span class="sd"> `torch.nn.functional.max_unpool2d`.</span>
-<span class="sd"> """</span>
- <span class="n">x</span><span class="p">,</span> <span class="n">indices</span> <span class="o">=</span> <span class="n">F</span><span class="o">.</span><span class="n">max_pool2d</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">kernel_size</span><span class="o">=</span><span class="mi">2</span><span class="p">,</span> <span class="n">return_indices</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
- <span class="k">return</span> <span class="n">x</span><span class="p">,</span> <span class="n">indices</span></div>
-
-<div class="viewcode-block" id="ConvBnReluMaxPool.extra_repr"><a class="viewcode-back" href="../../source/core.html#core.layers.ConvBnReluMaxPool.extra_repr">[docs]</a> <span class="k">def</span> <span class="nf">extra_repr</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Define optional extra information about this module.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> `str`</span>
-<span class="sd"> Extra representation string.</span>
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="p">(</span><span class="s1">'(pool): MaxPool2d(kernel_size=2, stride=2, padding=0, '</span>
- <span class="s1">'dilation=1, ceil_mode=False)'</span><span class="p">)</span></div></div>
-
-
-<div class="viewcode-block" id="ConvBnReluMaxUnpool"><a class="viewcode-back" href="../../source/core.html#core.layers.ConvBnReluMaxUnpool">[docs]</a><span class="k">class</span> <span class="nc">ConvBnReluMaxUnpool</span><span class="p">(</span><span class="n">DecoderBlock</span><span class="p">):</span>
- <span class="sd">"""Block of convolution, batchnorm, relu and 2x2 max unpool.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> in_channels : `int`</span>
-<span class="sd"> Number of input channels.</span>
-<span class="sd"> out_channels : `int`</span>
-<span class="sd"> Number of output channels</span>
-<span class="sd"> **kwargs: 'dict' [`str`]</span>
-<span class="sd"> Additional keyword arguments passed to</span>
-<span class="sd"> `pysegcnn.core.layers.Conv2dSame`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">in_channels</span><span class="p">,</span> <span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">in_channels</span><span class="p">,</span> <span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span>
-
-<div class="viewcode-block" id="ConvBnReluMaxUnpool.layers"><a class="viewcode-back" href="../../source/core.html#core.layers.ConvBnReluMaxUnpool.layers">[docs]</a> <span class="k">def</span> <span class="nf">layers</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Sequence of convolution, batchnorm and relu layers.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> layers : `torch.nn.Sequential` [`torch.nn.Module`]</span>
-<span class="sd"> An instance of `torch.nn.Sequential` containing the sequence</span>
-<span class="sd"> of convolution, batchnorm and relu layer (`torch.nn.Module`)</span>
-<span class="sd"> instances.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="n">conv_bn_relu</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">in_channels</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="bp">self</span><span class="o">.</span><span class="n">kwargs</span><span class="p">)</span></div>
-
-<div class="viewcode-block" id="ConvBnReluMaxUnpool.upsample"><a class="viewcode-back" href="../../source/core.html#core.layers.ConvBnReluMaxUnpool.upsample">[docs]</a> <span class="k">def</span> <span class="nf">upsample</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">x</span><span class="p">,</span> <span class="n">feature</span><span class="p">,</span> <span class="n">indices</span><span class="p">):</span>
- <span class="sd">"""2x2 max unpooling layer.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
-<span class="sd"> Input tensor.</span>
-<span class="sd"> feature : `torch.Tensor`, shape=(batch, channel, height, width)</span>
-<span class="sd"> Intermediate output of a layer in the encoder. Used to determine</span>
-<span class="sd"> the output shape of the upsampling operation.</span>
-<span class="sd"> indices : `torch.Tensor`</span>
-<span class="sd"> The indices of the maxima of the max pooling operation</span>
-<span class="sd"> (as returned by `torch.nn.functional.max_pool2d`).</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height * 2, width * 2)</span>
-<span class="sd"> The 2x2 max unpooled tensor.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="n">F</span><span class="o">.</span><span class="n">max_unpool2d</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">indices</span><span class="p">,</span> <span class="n">kernel_size</span><span class="o">=</span><span class="mi">2</span><span class="p">,</span>
- <span class="n">output_size</span><span class="o">=</span><span class="n">feature</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">2</span><span class="p">:])</span></div>
-
-<div class="viewcode-block" id="ConvBnReluMaxUnpool.extra_repr"><a class="viewcode-back" href="../../source/core.html#core.layers.ConvBnReluMaxUnpool.extra_repr">[docs]</a> <span class="k">def</span> <span class="nf">extra_repr</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Define optional extra information about this module.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> `str`</span>
-<span class="sd"> Extra representation string.</span>
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="p">(</span><span class="s1">'(pool): MaxUnpool2d(kernel_size=(2, 2), stride=(2, 2), '</span>
- <span class="s1">'padding=(0, 0))'</span><span class="p">)</span></div></div>
-
-
-<div class="viewcode-block" id="ConvBnReluUpsample"><a class="viewcode-back" href="../../source/core.html#core.layers.ConvBnReluUpsample">[docs]</a><span class="k">class</span> <span class="nc">ConvBnReluUpsample</span><span class="p">(</span><span class="n">DecoderBlock</span><span class="p">):</span>
- <span class="sd">"""Block of convolution, batchnorm, relu and nearest neighbor upsampling.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> in_channels : `int`</span>
-<span class="sd"> Number of input channels.</span>
-<span class="sd"> out_channels : `int`</span>
-<span class="sd"> Number of output channels</span>
-<span class="sd"> **kwargs: 'dict' [`str`]</span>
-<span class="sd"> Additional arguments passed to `pysegcnn.core.layers.Conv2dSame`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">in_channels</span><span class="p">,</span> <span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">in_channels</span><span class="p">,</span> <span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span>
-
-<div class="viewcode-block" id="ConvBnReluUpsample.layers"><a class="viewcode-back" href="../../source/core.html#core.layers.ConvBnReluUpsample.layers">[docs]</a> <span class="k">def</span> <span class="nf">layers</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Sequence of convolution, batchnorm and relu layers.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> layers : `torch.nn.Sequential` [`torch.nn.Module`]</span>
-<span class="sd"> An instance of `torch.nn.Sequential` containing the sequence</span>
-<span class="sd"> of convolution, batchnorm and relu layer (`torch.nn.Module`)</span>
-<span class="sd"> instances.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="n">conv_bn_relu</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">in_channels</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="bp">self</span><span class="o">.</span><span class="n">kwargs</span><span class="p">)</span></div>
-
-<div class="viewcode-block" id="ConvBnReluUpsample.upsample"><a class="viewcode-back" href="../../source/core.html#core.layers.ConvBnReluUpsample.upsample">[docs]</a> <span class="k">def</span> <span class="nf">upsample</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">x</span><span class="p">,</span> <span class="n">feature</span><span class="p">,</span> <span class="n">indices</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
- <span class="sd">"""Nearest neighbor upsampling.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
-<span class="sd"> Input tensor.</span>
-<span class="sd"> feature : `torch.Tensor`, shape=(batch, channel, height, width)</span>
-<span class="sd"> Intermediate output of a layer in the encoder. Used to determine</span>
-<span class="sd"> the output shape of the upsampling operation.</span>
-<span class="sd"> indices : `None`, optional</span>
-<span class="sd"> The indices of the maxima of the max pooling operation</span>
-<span class="sd"> (as returned by `torch.nn.functional.max_pool2d`). Not required by</span>
-<span class="sd"> this upsampling method.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
-<span class="sd"> The 2x2 upsampled tensor.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="n">F</span><span class="o">.</span><span class="n">interpolate</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">size</span><span class="o">=</span><span class="n">feature</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">2</span><span class="p">:],</span> <span class="n">mode</span><span class="o">=</span><span class="s1">'nearest'</span><span class="p">)</span></div>
-
-<div class="viewcode-block" id="ConvBnReluUpsample.extra_repr"><a class="viewcode-back" href="../../source/core.html#core.layers.ConvBnReluUpsample.extra_repr">[docs]</a> <span class="k">def</span> <span class="nf">extra_repr</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Define optional extra information about this module.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> `str`</span>
-<span class="sd"> Extra representation string.</span>
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="s1">'(pool): Upsample(mode="nearest")'</span></div></div>
-</pre></div>
-
- </div>
-
- </div>
- <footer>
-
-
- <hr/>
-
- <div role="contentinfo">
- <p>
-
- © Copyright 2020, Daniel Frisinghelli
-
- </p>
- </div>
-
-
-
- Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a
-
- <a href="https://github.com/rtfd/sphinx_rtd_theme">theme</a>
-
- provided by <a href="https://readthedocs.org">Read the Docs</a>.
-
-</footer>
-
- </div>
- </div>
-
- </section>
-
- </div>
-
-
- <script type="text/javascript">
- jQuery(function () {
- SphinxRtdTheme.Navigation.enable(true);
- });
- </script>
-
-
-
-
-
-
-</body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/_modules/core/logging.html b/docs/_build/html/_modules/core/logging.html
deleted file mode 100644
index 11127c5d2986e6d074a7d8ceb2514b0173f85232..0000000000000000000000000000000000000000
--- a/docs/_build/html/_modules/core/logging.html
+++ /dev/null
@@ -1,282 +0,0 @@
-
-
-<!DOCTYPE html>
-<html class="writer-html5" lang="en" >
-<head>
- <meta charset="utf-8">
-
- <meta name="viewport" content="width=device-width, initial-scale=1.0">
-
- <title>core.logging — PySegCNN documentation</title>
-
-
-
- <link rel="stylesheet" href="../../_static/css/theme.css" type="text/css" />
- <link rel="stylesheet" href="../../_static/pygments.css" type="text/css" />
-
-
-
-
-
-
-
- <!--[if lt IE 9]>
- <script src="../../_static/js/html5shiv.min.js"></script>
- <![endif]-->
-
-
- <script type="text/javascript" id="documentation_options" data-url_root="../../" src="../../_static/documentation_options.js"></script>
- <script src="../../_static/jquery.js"></script>
- <script src="../../_static/underscore.js"></script>
- <script src="../../_static/doctools.js"></script>
- <script src="../../_static/language_data.js"></script>
-
- <script type="text/javascript" src="../../_static/js/theme.js"></script>
-
-
- <link rel="index" title="Index" href="../../genindex.html" />
- <link rel="search" title="Search" href="../../search.html" />
-</head>
-
-<body class="wy-body-for-nav">
-
-
- <div class="wy-grid-for-nav">
-
- <nav data-toggle="wy-nav-shift" class="wy-nav-side">
- <div class="wy-side-scroll">
- <div class="wy-side-nav-search" >
-
-
-
- <a href="../../index.html" class="icon icon-home" alt="Documentation Home"> PySegCNN
-
-
-
- </a>
-
-
-
-
-
-
-
-<div role="search">
- <form id="rtd-search-form" class="wy-form" action="../../search.html" method="get">
- <input type="text" name="q" placeholder="Search docs" />
- <input type="hidden" name="check_keywords" value="yes" />
- <input type="hidden" name="area" value="default" />
- </form>
-</div>
-
-
- </div>
-
-
- <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
-
-
-
-
-
-
- <ul>
-<li class="toctree-l1"><a class="reference internal" href="../../source/requirements.html">Requirements</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../../source/installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../../source/api.html">API Reference</a></li>
-</ul>
-
-
-
- </div>
-
- </div>
- </nav>
-
- <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
-
-
- <nav class="wy-nav-top" aria-label="top navigation">
-
- <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
- <a href="../../index.html">PySegCNN</a>
-
- </nav>
-
-
- <div class="wy-nav-content">
-
- <div class="rst-content">
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-<div role="navigation" aria-label="breadcrumbs navigation">
-
- <ul class="wy-breadcrumbs">
-
- <li><a href="../../index.html" class="icon icon-home"></a> »</li>
-
- <li><a href="../index.html">Module code</a> »</li>
-
- <li>core.logging</li>
-
-
- <li class="wy-breadcrumbs-aside">
-
- </li>
-
- </ul>
-
-
- <hr/>
-</div>
- <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
- <div itemprop="articleBody">
-
- <h1>Source code for core.logging</h1><div class="highlight"><pre>
-<span></span><span class="sd">"""Logging configuration.</span>
-
-<span class="sd">License</span>
-<span class="sd">-------</span>
-
-<span class="sd"> Copyright (c) 2020 Daniel Frisinghelli</span>
-
-<span class="sd"> This source code is licensed under the GNU General Public License v3.</span>
-
-<span class="sd"> See the LICENSE file in the repository's root directory.</span>
-
-<span class="sd">"""</span>
-
-<span class="c1"># !/usr/bin/env python</span>
-<span class="c1"># -*- coding: utf-8 -*-</span>
-
-<span class="c1"># builtins</span>
-<span class="kn">import</span> <span class="nn">pathlib</span>
-
-
-<span class="c1"># the logging configuration dictionary</span>
-<div class="viewcode-block" id="log_conf"><a class="viewcode-back" href="../../source/core.html#core.logging.log_conf">[docs]</a><span class="k">def</span> <span class="nf">log_conf</span><span class="p">(</span><span class="n">logfile</span><span class="p">):</span>
- <span class="sd">"""Set basic logging configuration passed to `logging.config.dictConfig`.</span>
-
-<span class="sd"> See the logging `docs`_ for a detailed description of the configuration</span>
-<span class="sd"> dictionary.</span>
-
-<span class="sd"> .. _docs:</span>
-<span class="sd"> https://docs.python.org/3/library/logging.config.html#dictionary-schema-details</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> logfile : `str` or `pathlib.Path`</span>
-<span class="sd"> The file to save the logs to.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> LOGGING_CONFIG : `dict`</span>
-<span class="sd"> The logging configuration.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># check if the parent directory of the log file exists</span>
- <span class="n">logfile</span> <span class="o">=</span> <span class="n">pathlib</span><span class="o">.</span><span class="n">Path</span><span class="p">(</span><span class="n">logfile</span><span class="p">)</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="n">logfile</span><span class="o">.</span><span class="n">parent</span><span class="o">.</span><span class="n">is_dir</span><span class="p">():</span>
- <span class="n">logfile</span><span class="o">.</span><span class="n">parent</span><span class="o">.</span><span class="n">mkdir</span><span class="p">(</span><span class="n">parents</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span> <span class="n">exist_ok</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
-
- <span class="n">LOGGING_CONFIG</span> <span class="o">=</span> <span class="p">{</span>
- <span class="s1">'version'</span><span class="p">:</span> <span class="mi">1</span><span class="p">,</span>
- <span class="s1">'disable_existing_loggers'</span><span class="p">:</span> <span class="kc">False</span><span class="p">,</span>
- <span class="s1">'formatters'</span><span class="p">:</span> <span class="p">{</span>
- <span class="s1">'brief'</span><span class="p">:</span> <span class="p">{</span>
- <span class="s1">'format'</span><span class="p">:</span> <span class="s1">'</span><span class="si">%(name)s</span><span class="s1">: </span><span class="si">%(message)s</span><span class="s1">'</span>
- <span class="p">},</span>
- <span class="s1">'standard'</span><span class="p">:</span> <span class="p">{</span>
- <span class="s1">'format'</span><span class="p">:</span> <span class="s1">'</span><span class="si">%(asctime)s</span><span class="s1"> [</span><span class="si">%(levelname)s</span><span class="s1">] </span><span class="si">%(name)s</span><span class="s1">: </span><span class="si">%(message)s</span><span class="s1">'</span><span class="p">,</span>
- <span class="s1">'datefmt'</span><span class="p">:</span> <span class="s1">'%Y-%m-</span><span class="si">%d</span><span class="s1">T%H:%M:%S'</span>
- <span class="p">},</span>
- <span class="p">},</span>
- <span class="s1">'handlers'</span><span class="p">:</span> <span class="p">{</span>
- <span class="s1">'console'</span><span class="p">:</span> <span class="p">{</span>
- <span class="s1">'class'</span><span class="p">:</span> <span class="s1">'logging.StreamHandler'</span><span class="p">,</span>
- <span class="s1">'formatter'</span><span class="p">:</span> <span class="s1">'brief'</span><span class="p">,</span>
- <span class="s1">'level'</span><span class="p">:</span> <span class="s1">'INFO'</span><span class="p">,</span>
- <span class="s1">'stream'</span><span class="p">:</span> <span class="s1">'ext://sys.stderr'</span><span class="p">,</span>
- <span class="p">},</span>
-
- <span class="s1">'file'</span><span class="p">:</span> <span class="p">{</span>
- <span class="s1">'class'</span><span class="p">:</span> <span class="s1">'logging.FileHandler'</span><span class="p">,</span>
- <span class="s1">'formatter'</span><span class="p">:</span> <span class="s1">'standard'</span><span class="p">,</span>
- <span class="s1">'level'</span><span class="p">:</span> <span class="s1">'INFO'</span><span class="p">,</span>
- <span class="s1">'filename'</span><span class="p">:</span> <span class="n">logfile</span><span class="p">,</span>
- <span class="s1">'mode'</span><span class="p">:</span> <span class="s1">'a'</span>
- <span class="p">}</span>
- <span class="p">},</span>
- <span class="s1">'loggers'</span><span class="p">:</span> <span class="p">{</span>
- <span class="s1">''</span><span class="p">:</span> <span class="p">{</span>
- <span class="s1">'handlers'</span><span class="p">:</span> <span class="p">[</span><span class="s1">'console'</span><span class="p">,</span> <span class="s1">'file'</span><span class="p">],</span>
- <span class="s1">'level'</span><span class="p">:</span> <span class="s1">'INFO'</span><span class="p">,</span>
- <span class="p">},</span>
- <span class="p">}</span>
- <span class="p">}</span>
-
- <span class="k">return</span> <span class="n">LOGGING_CONFIG</span></div>
-</pre></div>
-
- </div>
-
- </div>
- <footer>
-
-
- <hr/>
-
- <div role="contentinfo">
- <p>
-
- © Copyright 2020, Daniel Frisinghelli
-
- </p>
- </div>
-
-
-
- Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a
-
- <a href="https://github.com/rtfd/sphinx_rtd_theme">theme</a>
-
- provided by <a href="https://readthedocs.org">Read the Docs</a>.
-
-</footer>
-
- </div>
- </div>
-
- </section>
-
- </div>
-
-
- <script type="text/javascript">
- jQuery(function () {
- SphinxRtdTheme.Navigation.enable(true);
- });
- </script>
-
-
-
-
-
-
-</body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/_modules/core/models.html b/docs/_build/html/_modules/core/models.html
deleted file mode 100644
index a9f1a461bf768dbe89819e0b9377f6d25accf7b8..0000000000000000000000000000000000000000
--- a/docs/_build/html/_modules/core/models.html
+++ /dev/null
@@ -1,532 +0,0 @@
-
-
-<!DOCTYPE html>
-<html class="writer-html5" lang="en" >
-<head>
- <meta charset="utf-8">
-
- <meta name="viewport" content="width=device-width, initial-scale=1.0">
-
- <title>core.models — PySegCNN documentation</title>
-
-
-
- <link rel="stylesheet" href="../../_static/css/theme.css" type="text/css" />
- <link rel="stylesheet" href="../../_static/pygments.css" type="text/css" />
-
-
-
-
-
-
-
- <!--[if lt IE 9]>
- <script src="../../_static/js/html5shiv.min.js"></script>
- <![endif]-->
-
-
- <script type="text/javascript" id="documentation_options" data-url_root="../../" src="../../_static/documentation_options.js"></script>
- <script src="../../_static/jquery.js"></script>
- <script src="../../_static/underscore.js"></script>
- <script src="../../_static/doctools.js"></script>
- <script src="../../_static/language_data.js"></script>
-
- <script type="text/javascript" src="../../_static/js/theme.js"></script>
-
-
- <link rel="index" title="Index" href="../../genindex.html" />
- <link rel="search" title="Search" href="../../search.html" />
-</head>
-
-<body class="wy-body-for-nav">
-
-
- <div class="wy-grid-for-nav">
-
- <nav data-toggle="wy-nav-shift" class="wy-nav-side">
- <div class="wy-side-scroll">
- <div class="wy-side-nav-search" >
-
-
-
- <a href="../../index.html" class="icon icon-home" alt="Documentation Home"> PySegCNN
-
-
-
- </a>
-
-
-
-
-
-
-
-<div role="search">
- <form id="rtd-search-form" class="wy-form" action="../../search.html" method="get">
- <input type="text" name="q" placeholder="Search docs" />
- <input type="hidden" name="check_keywords" value="yes" />
- <input type="hidden" name="area" value="default" />
- </form>
-</div>
-
-
- </div>
-
-
- <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
-
-
-
-
-
-
- <ul>
-<li class="toctree-l1"><a class="reference internal" href="../../source/requirements.html">Requirements</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../../source/installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../../source/api.html">API Reference</a></li>
-</ul>
-
-
-
- </div>
-
- </div>
- </nav>
-
- <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
-
-
- <nav class="wy-nav-top" aria-label="top navigation">
-
- <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
- <a href="../../index.html">PySegCNN</a>
-
- </nav>
-
-
- <div class="wy-nav-content">
-
- <div class="rst-content">
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-<div role="navigation" aria-label="breadcrumbs navigation">
-
- <ul class="wy-breadcrumbs">
-
- <li><a href="../../index.html" class="icon icon-home"></a> »</li>
-
- <li><a href="../index.html">Module code</a> »</li>
-
- <li>core.models</li>
-
-
- <li class="wy-breadcrumbs-aside">
-
- </li>
-
- </ul>
-
-
- <hr/>
-</div>
- <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
- <div itemprop="articleBody">
-
- <h1>Source code for core.models</h1><div class="highlight"><pre>
-<span></span><span class="sd">"""Neural networks for semantic image segmentation.</span>
-
-<span class="sd">License</span>
-<span class="sd">-------</span>
-
-<span class="sd"> Copyright (c) 2020 Daniel Frisinghelli</span>
-
-<span class="sd"> This source code is licensed under the GNU General Public License v3.</span>
-
-<span class="sd"> See the LICENSE file in the repository's root directory.</span>
-
-<span class="sd">"""</span>
-
-<span class="c1"># !/usr/bin/env python</span>
-<span class="c1"># -*- coding: utf-8 -*-</span>
-
-<span class="c1"># builtins</span>
-<span class="kn">import</span> <span class="nn">enum</span>
-<span class="kn">import</span> <span class="nn">logging</span>
-<span class="kn">import</span> <span class="nn">pathlib</span>
-
-<span class="c1"># externals</span>
-<span class="kn">import</span> <span class="nn">numpy</span> <span class="k">as</span> <span class="nn">np</span>
-<span class="kn">import</span> <span class="nn">torch</span>
-<span class="kn">import</span> <span class="nn">torch.nn</span> <span class="k">as</span> <span class="nn">nn</span>
-<span class="kn">import</span> <span class="nn">torch.optim</span> <span class="k">as</span> <span class="nn">optim</span>
-
-<span class="c1"># locals</span>
-<span class="kn">from</span> <span class="nn">pysegcnn.core.layers</span> <span class="kn">import</span> <span class="p">(</span><span class="n">Encoder</span><span class="p">,</span> <span class="n">Decoder</span><span class="p">,</span> <span class="n">ConvBnReluMaxPool</span><span class="p">,</span>
- <span class="n">ConvBnReluMaxUnpool</span><span class="p">,</span> <span class="n">Conv2dSame</span><span class="p">)</span>
-
-<span class="c1"># module level logger</span>
-<span class="n">LOGGER</span> <span class="o">=</span> <span class="n">logging</span><span class="o">.</span><span class="n">getLogger</span><span class="p">(</span><span class="vm">__name__</span><span class="p">)</span>
-
-
-<div class="viewcode-block" id="Network"><a class="viewcode-back" href="../../source/core.html#core.models.Network">[docs]</a><span class="k">class</span> <span class="nc">Network</span><span class="p">(</span><span class="n">nn</span><span class="o">.</span><span class="n">Module</span><span class="p">):</span>
- <span class="sd">"""Generic Network class.</span>
-
-<span class="sd"> The base class for each model. If you want to implement a new model,</span>
-<span class="sd"> inherit the ``~pysegcnn.core.models.Network`` class.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">()</span>
-
- <span class="c1"># initialize state file</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">state_file</span> <span class="o">=</span> <span class="kc">None</span>
-
-<div class="viewcode-block" id="Network.freeze"><a class="viewcode-back" href="../../source/core.html#core.models.Network.freeze">[docs]</a> <span class="k">def</span> <span class="nf">freeze</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Freeze the weights of a model.</span>
-
-<span class="sd"> Disables gradient computation: useful when using a pretrained model for</span>
-<span class="sd"> inference.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
- <span class="k">for</span> <span class="n">param</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">parameters</span><span class="p">():</span>
- <span class="n">param</span><span class="o">.</span><span class="n">requires_grad</span> <span class="o">=</span> <span class="kc">False</span></div>
-
-<div class="viewcode-block" id="Network.unfreeze"><a class="viewcode-back" href="../../source/core.html#core.models.Network.unfreeze">[docs]</a> <span class="k">def</span> <span class="nf">unfreeze</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Unfreeze the weights of a model.</span>
-
-<span class="sd"> Enables gradient computation: useful when adjusting a pretrained model</span>
-<span class="sd"> to a new dataset.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
- <span class="k">for</span> <span class="n">param</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">parameters</span><span class="p">():</span>
- <span class="n">param</span><span class="o">.</span><span class="n">requires_grad</span> <span class="o">=</span> <span class="kc">True</span></div>
-
-<div class="viewcode-block" id="Network.save"><a class="viewcode-back" href="../../source/core.html#core.models.Network.save">[docs]</a> <span class="k">def</span> <span class="nf">save</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">state_file</span><span class="p">,</span> <span class="n">optimizer</span><span class="p">,</span> <span class="n">bands</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
- <span class="sd">"""Save the model state.</span>
-
-<span class="sd"> Saves the model and optimizer states together with the model</span>
-<span class="sd"> construction parameters, to easily re-instanciate the model.</span>
-
-<span class="sd"> Optional ``kwargs`` are also saved.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> state_file : `str` or `pathlib.Path`</span>
-<span class="sd"> Path to save the model state.</span>
-<span class="sd"> optimizer : `torch.optim.Optimizer`</span>
-<span class="sd"> The optimizer used to train the model.</span>
-<span class="sd"> bands : `list` [`str`] or `None`, optional</span>
-<span class="sd"> List of bands the model is trained with. The default is None.</span>
-<span class="sd"> **kwargs</span>
-<span class="sd"> Arbitrary keyword arguments. Each keyword argument will be saved</span>
-<span class="sd"> as (key, value) pair in ``state_file``.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> model_state : `dict`</span>
-<span class="sd"> A dictionary containing the model and optimizer state</span>
-
-<span class="sd"> """</span>
- <span class="c1"># check if the output path exists and if not, create it</span>
- <span class="n">state_file</span> <span class="o">=</span> <span class="n">pathlib</span><span class="o">.</span><span class="n">Path</span><span class="p">(</span><span class="n">state_file</span><span class="p">)</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="n">state_file</span><span class="o">.</span><span class="n">parent</span><span class="o">.</span><span class="n">is_dir</span><span class="p">():</span>
- <span class="n">state_file</span><span class="o">.</span><span class="n">parent</span><span class="o">.</span><span class="n">mkdir</span><span class="p">(</span><span class="n">parents</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span> <span class="n">exist_ok</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
-
- <span class="c1"># initialize dictionary to store network parameters</span>
- <span class="n">model_state</span> <span class="o">=</span> <span class="p">{</span><span class="o">**</span><span class="n">kwargs</span><span class="p">}</span>
-
- <span class="c1"># store the spectral bands the model is trained with</span>
- <span class="n">model_state</span><span class="p">[</span><span class="s1">'bands'</span><span class="p">]</span> <span class="o">=</span> <span class="n">bands</span>
-
- <span class="c1"># store model and optimizer class</span>
- <span class="n">model_state</span><span class="p">[</span><span class="s1">'cls'</span><span class="p">]</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="vm">__class__</span>
- <span class="n">model_state</span><span class="p">[</span><span class="s1">'optim_cls'</span><span class="p">]</span> <span class="o">=</span> <span class="n">optimizer</span><span class="o">.</span><span class="vm">__class__</span>
-
- <span class="c1"># store construction parameters to instanciate the network</span>
- <span class="n">model_state</span><span class="p">[</span><span class="s1">'params'</span><span class="p">]</span> <span class="o">=</span> <span class="p">{</span>
- <span class="s1">'skip'</span><span class="p">:</span> <span class="bp">self</span><span class="o">.</span><span class="n">skip</span><span class="p">,</span>
- <span class="s1">'filters'</span><span class="p">:</span> <span class="bp">self</span><span class="o">.</span><span class="n">nfilters</span><span class="p">,</span>
- <span class="s1">'nclasses'</span><span class="p">:</span> <span class="bp">self</span><span class="o">.</span><span class="n">nclasses</span><span class="p">,</span>
- <span class="s1">'in_channels'</span><span class="p">:</span> <span class="bp">self</span><span class="o">.</span><span class="n">in_channels</span>
- <span class="p">}</span>
-
- <span class="c1"># store optimizer construction parameters</span>
- <span class="n">model_state</span><span class="p">[</span><span class="s1">'optim_params'</span><span class="p">]</span> <span class="o">=</span> <span class="n">optimizer</span><span class="o">.</span><span class="n">defaults</span>
-
- <span class="c1"># store optional keyword arguments</span>
- <span class="n">model_state</span><span class="p">[</span><span class="s1">'params'</span><span class="p">]</span> <span class="o">=</span> <span class="p">{</span><span class="o">**</span><span class="n">model_state</span><span class="p">[</span><span class="s1">'params'</span><span class="p">],</span> <span class="o">**</span><span class="bp">self</span><span class="o">.</span><span class="n">kwargs</span><span class="p">}</span>
-
- <span class="c1"># store model epoch</span>
- <span class="n">model_state</span><span class="p">[</span><span class="s1">'epoch'</span><span class="p">]</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">epoch</span>
-
- <span class="c1"># store model and optimizer state</span>
- <span class="n">model_state</span><span class="p">[</span><span class="s1">'model_state_dict'</span><span class="p">]</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">state_dict</span><span class="p">()</span>
- <span class="n">model_state</span><span class="p">[</span><span class="s1">'optim_state_dict'</span><span class="p">]</span> <span class="o">=</span> <span class="n">optimizer</span><span class="o">.</span><span class="n">state_dict</span><span class="p">()</span>
-
- <span class="c1"># model state dictionary stores the values of all trainable parameters</span>
- <span class="n">torch</span><span class="o">.</span><span class="n">save</span><span class="p">(</span><span class="n">model_state</span><span class="p">,</span> <span class="n">state_file</span><span class="p">)</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Network parameters saved in </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">state_file</span><span class="p">))</span>
-
- <span class="k">return</span> <span class="n">model_state</span></div>
-
-<div class="viewcode-block" id="Network.load"><a class="viewcode-back" href="../../source/core.html#core.models.Network.load">[docs]</a> <span class="nd">@staticmethod</span>
- <span class="k">def</span> <span class="nf">load</span><span class="p">(</span><span class="n">state_file</span><span class="p">):</span>
- <span class="sd">"""Load a model state.</span>
-
-<span class="sd"> Returns the model in ``state_file`` with the pretrained model and</span>
-<span class="sd"> optimizer weights. Useful when resuming training an existing model.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> state_file : `str` or `pathlib.Path`</span>
-<span class="sd"> The model state file. Model state files are stored in</span>
-<span class="sd"> pysegcnn/main/_models.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> FileNotFoundError</span>
-<span class="sd"> Raised if ``state_file`` does not exist.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> model : `pysegcnn.core.models.Network`</span>
-<span class="sd"> The pretrained model.</span>
-<span class="sd"> optimizer : `torch.optim.Optimizer`</span>
-<span class="sd"> The optimizer used to train the model.</span>
-<span class="sd"> model_state : '`dict`</span>
-<span class="sd"> A dictionary containing the model and optimizer state, as</span>
-<span class="sd"> constructed by `~pysegcnn.core.Network.save`.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># load the pretrained model</span>
- <span class="n">state_file</span> <span class="o">=</span> <span class="n">pathlib</span><span class="o">.</span><span class="n">Path</span><span class="p">(</span><span class="n">state_file</span><span class="p">)</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="n">state_file</span><span class="o">.</span><span class="n">exists</span><span class="p">():</span>
- <span class="k">raise</span> <span class="ne">FileNotFoundError</span><span class="p">(</span><span class="s1">'</span><span class="si">{}</span><span class="s1"> does not exist.'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">state_file</span><span class="p">))</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Loading pretrained weights from: </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">state_file</span><span class="p">))</span>
-
- <span class="c1"># load the model state</span>
- <span class="n">model_state</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">load</span><span class="p">(</span><span class="n">state_file</span><span class="p">)</span>
-
- <span class="c1"># the model and optimizer class</span>
- <span class="n">model_class</span> <span class="o">=</span> <span class="n">model_state</span><span class="p">[</span><span class="s1">'cls'</span><span class="p">]</span>
- <span class="n">optim_class</span> <span class="o">=</span> <span class="n">model_state</span><span class="p">[</span><span class="s1">'optim_cls'</span><span class="p">]</span>
-
- <span class="c1"># instanciate pretrained model architecture</span>
- <span class="n">model</span> <span class="o">=</span> <span class="n">model_class</span><span class="p">(</span><span class="o">**</span><span class="n">model_state</span><span class="p">[</span><span class="s1">'params'</span><span class="p">])</span>
-
- <span class="c1"># store state file as instance attribute</span>
- <span class="n">model</span><span class="o">.</span><span class="n">state_file</span> <span class="o">=</span> <span class="n">state_file</span>
-
- <span class="c1"># load pretrained model weights</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Loading model parameters ...'</span><span class="p">)</span>
- <span class="n">model</span><span class="o">.</span><span class="n">load_state_dict</span><span class="p">(</span><span class="n">model_state</span><span class="p">[</span><span class="s1">'model_state_dict'</span><span class="p">])</span>
- <span class="n">model</span><span class="o">.</span><span class="n">epoch</span> <span class="o">=</span> <span class="n">model_state</span><span class="p">[</span><span class="s1">'epoch'</span><span class="p">]</span>
-
- <span class="c1"># resume optimizer parameters</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Loading optimizer parameters ...'</span><span class="p">)</span>
- <span class="n">optimizer</span> <span class="o">=</span> <span class="n">optim_class</span><span class="p">(</span><span class="n">model</span><span class="o">.</span><span class="n">parameters</span><span class="p">(),</span>
- <span class="o">**</span><span class="n">model_state</span><span class="p">[</span><span class="s1">'optim_params'</span><span class="p">])</span>
- <span class="n">optimizer</span><span class="o">.</span><span class="n">load_state_dict</span><span class="p">(</span><span class="n">model_state</span><span class="p">[</span><span class="s1">'optim_state_dict'</span><span class="p">])</span>
-
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Model epoch: </span><span class="si">{:d}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">model</span><span class="o">.</span><span class="n">epoch</span><span class="p">))</span>
-
- <span class="k">return</span> <span class="n">model</span><span class="p">,</span> <span class="n">optimizer</span><span class="p">,</span> <span class="n">model_state</span></div>
-
- <span class="nd">@property</span>
- <span class="k">def</span> <span class="nf">state</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Return the model state file.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> state_file : `pathlib.Path` or `None`</span>
-<span class="sd"> The model state file.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">state_file</span></div>
-
-
-<div class="viewcode-block" id="UNet"><a class="viewcode-back" href="../../source/core.html#core.models.UNet">[docs]</a><span class="k">class</span> <span class="nc">UNet</span><span class="p">(</span><span class="n">Network</span><span class="p">):</span>
- <span class="sd">"""A PyTorch implementation of `U-Net`_.</span>
-
-<span class="sd"> Slightly modified version of U-Net:</span>
-<span class="sd"> - each convolution is followed by a batch normalization layer</span>
-<span class="sd"> - the upsampling is implemented by a 2x2 max unpooling operation</span>
-
-<span class="sd"> .. _U-Net:</span>
-<span class="sd"> https://arxiv.org/abs/1505.04597</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> in_channels : `int`</span>
-<span class="sd"> Number of channels of the input images.</span>
-<span class="sd"> nclasses : `int`</span>
-<span class="sd"> Number of classes.</span>
-<span class="sd"> filters : `list` [`int`]</span>
-<span class="sd"> List of input channels to each convolutional block.</span>
-<span class="sd"> skip : `bool`</span>
-<span class="sd"> Whether to apply skip connections from the encoder to the decoder.</span>
-<span class="sd"> **kwargs: 'dict' [`str`]</span>
-<span class="sd"> Additional keyword arguments passed to</span>
-<span class="sd"> `pysegcnn.core.layers.Conv2dSame`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">in_channels</span><span class="p">,</span> <span class="n">nclasses</span><span class="p">,</span> <span class="n">filters</span><span class="p">,</span> <span class="n">skip</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">()</span>
-
- <span class="c1"># number of input channels</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">in_channels</span> <span class="o">=</span> <span class="n">in_channels</span>
-
- <span class="c1"># number of classes</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">nclasses</span> <span class="o">=</span> <span class="n">nclasses</span>
-
- <span class="c1"># configuration of the convolutional layers in the network</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">kwargs</span> <span class="o">=</span> <span class="n">kwargs</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">nfilters</span> <span class="o">=</span> <span class="n">filters</span>
-
- <span class="c1"># convolutional layers of the encoder</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">filters</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">hstack</span><span class="p">([</span><span class="n">np</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">in_channels</span><span class="p">),</span> <span class="n">np</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">filters</span><span class="p">)])</span>
-
- <span class="c1"># whether to apply skip connections</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">skip</span> <span class="o">=</span> <span class="n">skip</span>
-
- <span class="c1"># number of epochs trained</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">epoch</span> <span class="o">=</span> <span class="mi">0</span>
-
- <span class="c1"># construct the encoder</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">encoder</span> <span class="o">=</span> <span class="n">Encoder</span><span class="p">(</span><span class="n">filters</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">filters</span><span class="p">,</span> <span class="n">block</span><span class="o">=</span><span class="n">ConvBnReluMaxPool</span><span class="p">,</span>
- <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span>
-
- <span class="c1"># construct the decoder</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">decoder</span> <span class="o">=</span> <span class="n">Decoder</span><span class="p">(</span><span class="n">filters</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">filters</span><span class="p">,</span> <span class="n">block</span><span class="o">=</span><span class="n">ConvBnReluMaxUnpool</span><span class="p">,</span>
- <span class="n">skip</span><span class="o">=</span><span class="n">skip</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span>
-
- <span class="c1"># construct the classifier</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">classifier</span> <span class="o">=</span> <span class="n">Conv2dSame</span><span class="p">(</span><span class="n">in_channels</span><span class="o">=</span><span class="n">filters</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span>
- <span class="n">out_channels</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">nclasses</span><span class="p">,</span>
- <span class="n">kernel_size</span><span class="o">=</span><span class="mi">1</span><span class="p">)</span>
-
-<div class="viewcode-block" id="UNet.forward"><a class="viewcode-back" href="../../source/core.html#core.models.UNet.forward">[docs]</a> <span class="k">def</span> <span class="nf">forward</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">x</span><span class="p">):</span>
- <span class="sd">"""Forward propagation of U-Net.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> x : `torch.Tensor`</span>
-<span class="sd"> The input image, shape=(batch_size, channels, height, width).</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> y : 'torch.tensor'</span>
-<span class="sd"> The classified image, shape=(batch_size, height, width).</span>
-
-<span class="sd"> """</span>
- <span class="c1"># forward pass: encoder</span>
- <span class="n">x</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">encoder</span><span class="p">(</span><span class="n">x</span><span class="p">)</span>
-
- <span class="c1"># forward pass: decoder</span>
- <span class="n">x</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">decoder</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">encoder</span><span class="o">.</span><span class="n">cache</span><span class="p">)</span>
-
- <span class="c1"># clear intermediate outputs</span>
- <span class="k">del</span> <span class="bp">self</span><span class="o">.</span><span class="n">encoder</span><span class="o">.</span><span class="n">cache</span>
-
- <span class="c1"># classification</span>
- <span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">classifier</span><span class="p">(</span><span class="n">x</span><span class="p">)</span></div></div>
-
-
-<div class="viewcode-block" id="SupportedModels"><a class="viewcode-back" href="../../source/core.html#core.models.SupportedModels">[docs]</a><span class="k">class</span> <span class="nc">SupportedModels</span><span class="p">(</span><span class="n">enum</span><span class="o">.</span><span class="n">Enum</span><span class="p">):</span>
- <span class="sd">"""Names and corresponding classes of the implemented models."""</span>
-
- <span class="n">Unet</span> <span class="o">=</span> <span class="n">UNet</span></div>
-
-
-<div class="viewcode-block" id="SupportedOptimizers"><a class="viewcode-back" href="../../source/core.html#core.models.SupportedOptimizers">[docs]</a><span class="k">class</span> <span class="nc">SupportedOptimizers</span><span class="p">(</span><span class="n">enum</span><span class="o">.</span><span class="n">Enum</span><span class="p">):</span>
- <span class="sd">"""Names and corresponding classes of the tested optimizers."""</span>
-
- <span class="n">Adam</span> <span class="o">=</span> <span class="n">optim</span><span class="o">.</span><span class="n">Adam</span></div>
-
-
-<div class="viewcode-block" id="SupportedLossFunctions"><a class="viewcode-back" href="../../source/core.html#core.models.SupportedLossFunctions">[docs]</a><span class="k">class</span> <span class="nc">SupportedLossFunctions</span><span class="p">(</span><span class="n">enum</span><span class="o">.</span><span class="n">Enum</span><span class="p">):</span>
- <span class="sd">"""Names and corresponding classes of the tested loss functions."""</span>
-
- <span class="n">CrossEntropy</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">CrossEntropyLoss</span></div>
-</pre></div>
-
- </div>
-
- </div>
- <footer>
-
-
- <hr/>
-
- <div role="contentinfo">
- <p>
-
- © Copyright 2020, Daniel Frisinghelli
-
- </p>
- </div>
-
-
-
- Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a
-
- <a href="https://github.com/rtfd/sphinx_rtd_theme">theme</a>
-
- provided by <a href="https://readthedocs.org">Read the Docs</a>.
-
-</footer>
-
- </div>
- </div>
-
- </section>
-
- </div>
-
-
- <script type="text/javascript">
- jQuery(function () {
- SphinxRtdTheme.Navigation.enable(true);
- });
- </script>
-
-
-
-
-
-
-</body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/_modules/core/predict.html b/docs/_build/html/_modules/core/predict.html
deleted file mode 100644
index b8c6d4862a2f94578daca6f468b7d6a722e3e764..0000000000000000000000000000000000000000
--- a/docs/_build/html/_modules/core/predict.html
+++ /dev/null
@@ -1,504 +0,0 @@
-
-
-<!DOCTYPE html>
-<html class="writer-html5" lang="en" >
-<head>
- <meta charset="utf-8">
-
- <meta name="viewport" content="width=device-width, initial-scale=1.0">
-
- <title>core.predict — PySegCNN documentation</title>
-
-
-
- <link rel="stylesheet" href="../../_static/css/theme.css" type="text/css" />
- <link rel="stylesheet" href="../../_static/pygments.css" type="text/css" />
-
-
-
-
-
-
-
- <!--[if lt IE 9]>
- <script src="../../_static/js/html5shiv.min.js"></script>
- <![endif]-->
-
-
- <script type="text/javascript" id="documentation_options" data-url_root="../../" src="../../_static/documentation_options.js"></script>
- <script src="../../_static/jquery.js"></script>
- <script src="../../_static/underscore.js"></script>
- <script src="../../_static/doctools.js"></script>
- <script src="../../_static/language_data.js"></script>
-
- <script type="text/javascript" src="../../_static/js/theme.js"></script>
-
-
- <link rel="index" title="Index" href="../../genindex.html" />
- <link rel="search" title="Search" href="../../search.html" />
-</head>
-
-<body class="wy-body-for-nav">
-
-
- <div class="wy-grid-for-nav">
-
- <nav data-toggle="wy-nav-shift" class="wy-nav-side">
- <div class="wy-side-scroll">
- <div class="wy-side-nav-search" >
-
-
-
- <a href="../../index.html" class="icon icon-home" alt="Documentation Home"> PySegCNN
-
-
-
- </a>
-
-
-
-
-
-
-
-<div role="search">
- <form id="rtd-search-form" class="wy-form" action="../../search.html" method="get">
- <input type="text" name="q" placeholder="Search docs" />
- <input type="hidden" name="check_keywords" value="yes" />
- <input type="hidden" name="area" value="default" />
- </form>
-</div>
-
-
- </div>
-
-
- <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
-
-
-
-
-
-
- <ul>
-<li class="toctree-l1"><a class="reference internal" href="../../source/requirements.html">Requirements</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../../source/installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../../source/api.html">API Reference</a></li>
-</ul>
-
-
-
- </div>
-
- </div>
- </nav>
-
- <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
-
-
- <nav class="wy-nav-top" aria-label="top navigation">
-
- <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
- <a href="../../index.html">PySegCNN</a>
-
- </nav>
-
-
- <div class="wy-nav-content">
-
- <div class="rst-content">
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-<div role="navigation" aria-label="breadcrumbs navigation">
-
- <ul class="wy-breadcrumbs">
-
- <li><a href="../../index.html" class="icon icon-home"></a> »</li>
-
- <li><a href="../index.html">Module code</a> »</li>
-
- <li>core.predict</li>
-
-
- <li class="wy-breadcrumbs-aside">
-
- </li>
-
- </ul>
-
-
- <hr/>
-</div>
- <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
- <div itemprop="articleBody">
-
- <h1>Source code for core.predict</h1><div class="highlight"><pre>
-<span></span><span class="sd">"""Functions for model inference.</span>
-
-<span class="sd">License</span>
-<span class="sd">-------</span>
-
-<span class="sd"> Copyright (c) 2020 Daniel Frisinghelli</span>
-
-<span class="sd"> This source code is licensed under the GNU General Public License v3.</span>
-
-<span class="sd"> See the LICENSE file in the repository's root directory.</span>
-
-<span class="sd">"""</span>
-
-<span class="c1"># !/usr/bin/env python</span>
-<span class="c1"># -*- coding: utf-8 -*-</span>
-
-<span class="c1"># builtins</span>
-<span class="kn">import</span> <span class="nn">logging</span>
-
-<span class="c1"># externals</span>
-<span class="kn">import</span> <span class="nn">numpy</span> <span class="k">as</span> <span class="nn">np</span>
-<span class="kn">import</span> <span class="nn">torch</span>
-<span class="kn">from</span> <span class="nn">torch.utils.data</span> <span class="kn">import</span> <span class="n">DataLoader</span>
-<span class="kn">from</span> <span class="nn">torch.utils.data.dataset</span> <span class="kn">import</span> <span class="n">Subset</span>
-<span class="kn">import</span> <span class="nn">torch.nn.functional</span> <span class="k">as</span> <span class="nn">F</span>
-
-<span class="c1"># locals</span>
-<span class="kn">from</span> <span class="nn">pysegcnn.core.utils</span> <span class="kn">import</span> <span class="n">reconstruct_scene</span><span class="p">,</span> <span class="n">accuracy_function</span>
-<span class="kn">from</span> <span class="nn">pysegcnn.core.graphics</span> <span class="kn">import</span> <span class="n">plot_sample</span>
-<span class="kn">from</span> <span class="nn">pysegcnn.core.split</span> <span class="kn">import</span> <span class="n">RandomSubset</span><span class="p">,</span> <span class="n">SceneSubset</span>
-
-<span class="c1"># module level logger</span>
-<span class="n">LOGGER</span> <span class="o">=</span> <span class="n">logging</span><span class="o">.</span><span class="n">getLogger</span><span class="p">(</span><span class="vm">__name__</span><span class="p">)</span>
-
-
-<span class="k">def</span> <span class="nf">_get_scene_tiles</span><span class="p">(</span><span class="n">ds</span><span class="p">,</span> <span class="n">scene_id</span><span class="p">):</span>
- <span class="sd">"""Return the tiles of the scene with id = ``scene_id``.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> ds : `pysegcnn.core.split.RandomSubset` or</span>
-<span class="sd"> `pysegcnn.core.split.SceneSubset`</span>
-<span class="sd"> An instance of `~pysegcnn.core.split.RandomSubset` or</span>
-<span class="sd"> `~pysegcnn.core.split.SceneSubset`.</span>
-<span class="sd"> scene_id : `str`</span>
-<span class="sd"> A valid scene identifier.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> ValueError</span>
-<span class="sd"> Raised if ``scene_id`` is not a valid scene identifier for the dataset</span>
-<span class="sd"> ``ds``.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> indices : `list` [`int`]</span>
-<span class="sd"> List of indices of the tiles from scene with id ``scene_id`` in ``ds``.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># check if the scene id is valid</span>
- <span class="n">scene_meta</span> <span class="o">=</span> <span class="n">ds</span><span class="o">.</span><span class="n">dataset</span><span class="o">.</span><span class="n">parse_scene_id</span><span class="p">(</span><span class="n">scene_id</span><span class="p">)</span>
- <span class="k">if</span> <span class="n">scene_meta</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
- <span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="s1">'</span><span class="si">{}</span><span class="s1"> is not a valid scene identifier'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">scene_id</span><span class="p">))</span>
-
- <span class="c1"># iterate over the scenes of the dataset</span>
- <span class="n">indices</span> <span class="o">=</span> <span class="p">[]</span>
- <span class="k">for</span> <span class="n">i</span><span class="p">,</span> <span class="n">scene</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">(</span><span class="n">ds</span><span class="o">.</span><span class="n">scenes</span><span class="p">):</span>
- <span class="c1"># if the scene id matches a given id, save the index of the scene</span>
- <span class="k">if</span> <span class="n">scene</span><span class="p">[</span><span class="s1">'id'</span><span class="p">]</span> <span class="o">==</span> <span class="n">scene_id</span><span class="p">:</span>
- <span class="n">indices</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">i</span><span class="p">)</span>
-
- <span class="k">return</span> <span class="n">indices</span>
-
-
-<div class="viewcode-block" id="predict_samples"><a class="viewcode-back" href="../../source/core.html#core.predict.predict_samples">[docs]</a><span class="k">def</span> <span class="nf">predict_samples</span><span class="p">(</span><span class="n">ds</span><span class="p">,</span> <span class="n">model</span><span class="p">,</span> <span class="n">cm</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">plot</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
- <span class="sd">"""Classify each sample in ``ds`` with model ``model``.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> ds : `pysegcnn.core.split.RandomSubset` or</span>
-<span class="sd"> `pysegcnn.core.split.SceneSubset`</span>
-<span class="sd"> An instance of `~pysegcnn.core.split.RandomSubset` or</span>
-<span class="sd"> `~pysegcnn.core.split.SceneSubset`.</span>
-<span class="sd"> model : `pysegcnn.core.models.Network`</span>
-<span class="sd"> An instance of `~pysegcnn.core.models.Network`.</span>
-<span class="sd"> cm : `bool`, optional</span>
-<span class="sd"> Whether to compute the confusion matrix. The default is False.</span>
-<span class="sd"> plot : `bool`, optional</span>
-<span class="sd"> Whether to plot a false color composite, ground truth and model</span>
-<span class="sd"> prediction for each sample. The default is False.</span>
-<span class="sd"> **kwargs</span>
-<span class="sd"> Additional keyword arguments passed to</span>
-<span class="sd"> `pysegcnn.core.graphics.plot_sample`.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> TypeError</span>
-<span class="sd"> Raised if ``ds`` is not an instance of</span>
-<span class="sd"> `~pysegcnn.core.split.RandomSubset` or</span>
-<span class="sd"> `~pysegcnn.core.split.SceneSubset`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> output : `dict`</span>
-<span class="sd"> Output dictionary with keys:</span>
-<span class="sd"> ``'input'``</span>
-<span class="sd"> Model input data</span>
-<span class="sd"> ``'labels'``</span>
-<span class="sd"> The ground truth</span>
-<span class="sd"> ``'prediction'``</span>
-<span class="sd"> Model prediction</span>
-<span class="sd"> conf_mat : `numpy.ndarray`</span>
-<span class="sd"> The confusion matrix. Note that the confusion matrix ``conf_mat`` is</span>
-<span class="sd"> only computed if ``cm`` = True.</span>
-<span class="sd"> """</span>
- <span class="c1"># check whether the dataset is a valid subset, i.e.</span>
- <span class="c1"># an instance of pysegcnn.core.split.SceneSubset or</span>
- <span class="c1"># an instance of pysegcnn.core.split.RandomSubset</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">ds</span><span class="p">,</span> <span class="n">RandomSubset</span><span class="p">)</span> <span class="ow">or</span> <span class="ow">not</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">ds</span><span class="p">,</span> <span class="n">SceneSubset</span><span class="p">):</span>
- <span class="k">raise</span> <span class="ne">TypeError</span><span class="p">(</span><span class="s1">'ds should be an instance of </span><span class="si">{}</span><span class="s1"> or of </span><span class="si">{}</span><span class="s1">.'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="nb">repr</span><span class="p">(</span><span class="n">RandomSubset</span><span class="p">),</span> <span class="nb">repr</span><span class="p">(</span><span class="n">SceneSubset</span><span class="p">)))</span>
-
- <span class="c1"># the device to compute on, use gpu if available</span>
- <span class="n">device</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">device</span><span class="p">(</span><span class="s2">"cuda:0"</span> <span class="k">if</span> <span class="n">torch</span><span class="o">.</span><span class="n">cuda</span><span class="o">.</span><span class="n">is_available</span><span class="p">()</span> <span class="k">else</span> <span class="s2">"cpu"</span><span class="p">)</span>
-
- <span class="c1"># set the model to evaluation mode</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Setting model to evaluation mode ...'</span><span class="p">)</span>
- <span class="n">model</span><span class="o">.</span><span class="n">eval</span><span class="p">()</span>
- <span class="n">model</span><span class="o">.</span><span class="n">to</span><span class="p">(</span><span class="n">device</span><span class="p">)</span>
-
- <span class="c1"># base filename for each sample</span>
- <span class="n">fname</span> <span class="o">=</span> <span class="n">model</span><span class="o">.</span><span class="n">state_file</span><span class="o">.</span><span class="n">name</span><span class="o">.</span><span class="n">split</span><span class="p">(</span><span class="s1">'.pt'</span><span class="p">)[</span><span class="mi">0</span><span class="p">]</span>
-
- <span class="c1"># initialize confusion matrix</span>
- <span class="n">conf_mat</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="p">(</span><span class="n">model</span><span class="o">.</span><span class="n">nclasses</span><span class="p">,</span> <span class="n">model</span><span class="o">.</span><span class="n">nclasses</span><span class="p">))</span>
-
- <span class="c1"># create the dataloader</span>
- <span class="n">dataloader</span> <span class="o">=</span> <span class="n">DataLoader</span><span class="p">(</span><span class="n">ds</span><span class="p">,</span> <span class="n">batch_size</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">shuffle</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">drop_last</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
-
- <span class="c1"># iterate over the samples and plot inputs, ground truth and</span>
- <span class="c1"># model predictions</span>
- <span class="n">output</span> <span class="o">=</span> <span class="p">{}</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Predicting samples of the </span><span class="si">{}</span><span class="s1"> dataset ...'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">ds</span><span class="o">.</span><span class="n">name</span><span class="p">))</span>
- <span class="k">for</span> <span class="n">batch</span><span class="p">,</span> <span class="p">(</span><span class="n">inputs</span><span class="p">,</span> <span class="n">labels</span><span class="p">)</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">(</span><span class="n">dataloader</span><span class="p">):</span>
-
- <span class="c1"># send inputs and labels to device</span>
- <span class="n">inputs</span> <span class="o">=</span> <span class="n">inputs</span><span class="o">.</span><span class="n">to</span><span class="p">(</span><span class="n">device</span><span class="p">)</span>
- <span class="n">labels</span> <span class="o">=</span> <span class="n">labels</span><span class="o">.</span><span class="n">to</span><span class="p">(</span><span class="n">device</span><span class="p">)</span>
-
- <span class="c1"># compute model predictions</span>
- <span class="k">with</span> <span class="n">torch</span><span class="o">.</span><span class="n">no_grad</span><span class="p">():</span>
- <span class="n">prd</span> <span class="o">=</span> <span class="n">F</span><span class="o">.</span><span class="n">softmax</span><span class="p">(</span><span class="n">model</span><span class="p">(</span><span class="n">inputs</span><span class="p">),</span> <span class="n">dim</span><span class="o">=</span><span class="mi">1</span><span class="p">)</span><span class="o">.</span><span class="n">argmax</span><span class="p">(</span><span class="n">dim</span><span class="o">=</span><span class="mi">1</span><span class="p">)</span><span class="o">.</span><span class="n">squeeze</span><span class="p">()</span>
-
- <span class="c1"># store output for current batch</span>
- <span class="n">output</span><span class="p">[</span><span class="n">batch</span><span class="p">]</span> <span class="o">=</span> <span class="p">{</span><span class="s1">'input'</span><span class="p">:</span> <span class="n">inputs</span><span class="p">,</span> <span class="s1">'labels'</span><span class="p">:</span> <span class="n">labels</span><span class="p">,</span> <span class="s1">'prediction'</span><span class="p">:</span> <span class="n">prd</span><span class="p">}</span>
-
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Sample: </span><span class="si">{:d}</span><span class="s1">/</span><span class="si">{:d}</span><span class="s1">, Accuracy: </span><span class="si">{:.2f}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span>
- <span class="n">batch</span> <span class="o">+</span> <span class="mi">1</span><span class="p">,</span> <span class="nb">len</span><span class="p">(</span><span class="n">dataloader</span><span class="p">),</span> <span class="n">accuracy_function</span><span class="p">(</span><span class="n">prd</span><span class="p">,</span> <span class="n">labels</span><span class="p">)))</span>
-
- <span class="c1"># update confusion matrix</span>
- <span class="k">if</span> <span class="n">cm</span><span class="p">:</span>
- <span class="k">for</span> <span class="n">ytrue</span><span class="p">,</span> <span class="n">ypred</span> <span class="ow">in</span> <span class="nb">zip</span><span class="p">(</span><span class="n">labels</span><span class="o">.</span><span class="n">view</span><span class="p">(</span><span class="o">-</span><span class="mi">1</span><span class="p">),</span> <span class="n">prd</span><span class="o">.</span><span class="n">view</span><span class="p">(</span><span class="o">-</span><span class="mi">1</span><span class="p">)):</span>
- <span class="n">conf_mat</span><span class="p">[</span><span class="n">ytrue</span><span class="o">.</span><span class="n">long</span><span class="p">(),</span> <span class="n">ypred</span><span class="o">.</span><span class="n">long</span><span class="p">()]</span> <span class="o">+=</span> <span class="mi">1</span>
-
- <span class="c1"># save plot of current batch to disk</span>
- <span class="k">if</span> <span class="n">plot</span><span class="p">:</span>
-
- <span class="c1"># plot inputs, ground truth and model predictions</span>
- <span class="n">sname</span> <span class="o">=</span> <span class="n">fname</span> <span class="o">+</span> <span class="s1">'_</span><span class="si">{}</span><span class="s1">_</span><span class="si">{}</span><span class="s1">.pt'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">ds</span><span class="o">.</span><span class="n">name</span><span class="p">,</span> <span class="n">batch</span><span class="p">)</span>
- <span class="n">fig</span><span class="p">,</span> <span class="n">ax</span> <span class="o">=</span> <span class="n">plot_sample</span><span class="p">(</span><span class="n">inputs</span><span class="o">.</span><span class="n">numpy</span><span class="p">()</span><span class="o">.</span><span class="n">clip</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="mi">1</span><span class="p">),</span>
- <span class="n">ds</span><span class="o">.</span><span class="n">dataset</span><span class="o">.</span><span class="n">use_bands</span><span class="p">,</span>
- <span class="n">ds</span><span class="o">.</span><span class="n">dataset</span><span class="o">.</span><span class="n">labels</span><span class="p">,</span>
- <span class="n">y</span><span class="o">=</span><span class="n">labels</span><span class="p">,</span>
- <span class="n">y_pred</span><span class="o">=</span><span class="n">prd</span><span class="p">,</span>
- <span class="n">state</span><span class="o">=</span><span class="n">sname</span><span class="p">,</span>
- <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span>
-
- <span class="k">return</span> <span class="n">output</span><span class="p">,</span> <span class="n">conf_mat</span></div>
-
-
-<div class="viewcode-block" id="predict_scenes"><a class="viewcode-back" href="../../source/core.html#core.predict.predict_scenes">[docs]</a><span class="k">def</span> <span class="nf">predict_scenes</span><span class="p">(</span><span class="n">ds</span><span class="p">,</span> <span class="n">model</span><span class="p">,</span> <span class="n">scene_id</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">cm</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">plot</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
- <span class="sd">"""Classify each scene in ``ds`` with model ``model``.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> ds : `pysegcnn.core.split.SceneSubset`</span>
-<span class="sd"> An instance of `~pysegcnn.core.split.SceneSubset`.</span>
-<span class="sd"> model : `pysegcnn.core.models.Network`</span>
-<span class="sd"> An instance of `~pysegcnn.core.models.Network`.</span>
-<span class="sd"> scene_id : `str` or `None`</span>
-<span class="sd"> A valid scene identifier.</span>
-<span class="sd"> cm : `bool`, optional</span>
-<span class="sd"> Whether to compute the confusion matrix. The default is False.</span>
-<span class="sd"> plot : `bool`, optional</span>
-<span class="sd"> Whether to plot a false color composite, ground truth and model</span>
-<span class="sd"> prediction for each scene. The default is False.</span>
-<span class="sd"> **kwargs</span>
-<span class="sd"> Additional keyword arguments passed to</span>
-<span class="sd"> `pysegcnn.core.graphics.plot_sample`.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> TypeError</span>
-<span class="sd"> Raised if ``ds`` is not an instance of</span>
-<span class="sd"> `~pysegcnn.core.split.SceneSubset`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> output : `dict`</span>
-<span class="sd"> Output dictionary with keys:</span>
-<span class="sd"> ``'input'``</span>
-<span class="sd"> Model input data</span>
-<span class="sd"> ``'labels'``</span>
-<span class="sd"> The ground truth</span>
-<span class="sd"> ``'prediction'``</span>
-<span class="sd"> Model prediction</span>
-<span class="sd"> conf_mat : `numpy.ndarray`</span>
-<span class="sd"> The confusion matrix. Note that the confusion matrix ``conf_mat`` is</span>
-<span class="sd"> only computed if ``cm`` = True.</span>
-<span class="sd"> """</span>
- <span class="c1"># check whether the dataset is a valid subset, i.e. an instance of</span>
- <span class="c1"># pysegcnn.core.split.SceneSubset</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">ds</span><span class="p">,</span> <span class="n">SceneSubset</span><span class="p">):</span>
- <span class="k">raise</span> <span class="ne">TypeError</span><span class="p">(</span><span class="s1">'ds should be an instance of </span><span class="si">{}</span><span class="s1">.'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="nb">repr</span><span class="p">(</span><span class="n">SceneSubset</span><span class="p">)))</span>
-
- <span class="c1"># the device to compute on, use gpu if available</span>
- <span class="n">device</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">device</span><span class="p">(</span><span class="s2">"cuda:0"</span> <span class="k">if</span> <span class="n">torch</span><span class="o">.</span><span class="n">cuda</span><span class="o">.</span><span class="n">is_available</span><span class="p">()</span> <span class="k">else</span> <span class="s2">"cpu"</span><span class="p">)</span>
-
- <span class="c1"># set the model to evaluation mode</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Setting model to evaluation mode ...'</span><span class="p">)</span>
- <span class="n">model</span><span class="o">.</span><span class="n">eval</span><span class="p">()</span>
- <span class="n">model</span><span class="o">.</span><span class="n">to</span><span class="p">(</span><span class="n">device</span><span class="p">)</span>
-
- <span class="c1"># base filename for each scene</span>
- <span class="n">fname</span> <span class="o">=</span> <span class="n">model</span><span class="o">.</span><span class="n">state_file</span><span class="o">.</span><span class="n">name</span><span class="o">.</span><span class="n">split</span><span class="p">(</span><span class="s1">'.pt'</span><span class="p">)[</span><span class="mi">0</span><span class="p">]</span>
-
- <span class="c1"># initialize confusion matrix</span>
- <span class="n">conf_mat</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="p">(</span><span class="n">model</span><span class="o">.</span><span class="n">nclasses</span><span class="p">,</span> <span class="n">model</span><span class="o">.</span><span class="n">nclasses</span><span class="p">))</span>
-
- <span class="c1"># check whether a scene id is provided</span>
- <span class="k">if</span> <span class="n">scene_id</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
- <span class="n">scene_ids</span> <span class="o">=</span> <span class="n">ds</span><span class="o">.</span><span class="n">ids</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="c1"># the name of the selected scene</span>
- <span class="n">scene_ids</span> <span class="o">=</span> <span class="p">[</span><span class="n">scene_id</span><span class="p">]</span>
-
- <span class="c1"># iterate over the scenes</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Predicting scenes of the </span><span class="si">{}</span><span class="s1"> dataset ...'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">ds</span><span class="o">.</span><span class="n">name</span><span class="p">))</span>
- <span class="n">output</span> <span class="o">=</span> <span class="p">{}</span>
- <span class="k">for</span> <span class="n">i</span><span class="p">,</span> <span class="n">sid</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">(</span><span class="n">scene_ids</span><span class="p">):</span>
-
- <span class="c1"># filename for the current scene</span>
- <span class="n">sname</span> <span class="o">=</span> <span class="n">fname</span> <span class="o">+</span> <span class="s1">'_</span><span class="si">{}</span><span class="s1">_</span><span class="si">{}</span><span class="s1">.pt'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">ds</span><span class="o">.</span><span class="n">name</span><span class="p">,</span> <span class="n">sid</span><span class="p">)</span>
-
- <span class="c1"># get the indices of the tiles of the scene</span>
- <span class="n">indices</span> <span class="o">=</span> <span class="n">_get_scene_tiles</span><span class="p">(</span><span class="n">ds</span><span class="p">,</span> <span class="n">sid</span><span class="p">)</span>
- <span class="n">indices</span><span class="o">.</span><span class="n">sort</span><span class="p">()</span>
-
- <span class="c1"># create a subset of the dataset</span>
- <span class="n">scene_ds</span> <span class="o">=</span> <span class="n">Subset</span><span class="p">(</span><span class="n">ds</span><span class="p">,</span> <span class="n">indices</span><span class="p">)</span>
-
- <span class="c1"># create the dataloader</span>
- <span class="n">scene_dl</span> <span class="o">=</span> <span class="n">DataLoader</span><span class="p">(</span><span class="n">scene_ds</span><span class="p">,</span> <span class="n">batch_size</span><span class="o">=</span><span class="nb">len</span><span class="p">(</span><span class="n">scene_ds</span><span class="p">),</span>
- <span class="n">shuffle</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">drop_last</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
-
- <span class="c1"># predict the current scene</span>
- <span class="k">for</span> <span class="n">b</span><span class="p">,</span> <span class="p">(</span><span class="n">inp</span><span class="p">,</span> <span class="n">lab</span><span class="p">)</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">(</span><span class="n">scene_dl</span><span class="p">):</span>
-
- <span class="c1"># send inputs and labels to device</span>
- <span class="n">inp</span> <span class="o">=</span> <span class="n">inp</span><span class="o">.</span><span class="n">to</span><span class="p">(</span><span class="n">device</span><span class="p">)</span>
- <span class="n">lab</span> <span class="o">=</span> <span class="n">lab</span><span class="o">.</span><span class="n">to</span><span class="p">(</span><span class="n">device</span><span class="p">)</span>
-
- <span class="c1"># apply forward pass: model prediction</span>
- <span class="k">with</span> <span class="n">torch</span><span class="o">.</span><span class="n">no_grad</span><span class="p">():</span>
- <span class="n">prd</span> <span class="o">=</span> <span class="n">F</span><span class="o">.</span><span class="n">softmax</span><span class="p">(</span><span class="n">model</span><span class="p">(</span><span class="n">inp</span><span class="p">),</span> <span class="n">dim</span><span class="o">=</span><span class="mi">1</span><span class="p">)</span><span class="o">.</span><span class="n">argmax</span><span class="p">(</span><span class="n">dim</span><span class="o">=</span><span class="mi">1</span><span class="p">)</span><span class="o">.</span><span class="n">squeeze</span><span class="p">()</span>
-
- <span class="c1"># update confusion matrix</span>
- <span class="k">if</span> <span class="n">cm</span><span class="p">:</span>
- <span class="k">for</span> <span class="n">ytrue</span><span class="p">,</span> <span class="n">ypred</span> <span class="ow">in</span> <span class="nb">zip</span><span class="p">(</span><span class="n">lab</span><span class="o">.</span><span class="n">view</span><span class="p">(</span><span class="o">-</span><span class="mi">1</span><span class="p">),</span> <span class="n">prd</span><span class="o">.</span><span class="n">view</span><span class="p">(</span><span class="o">-</span><span class="mi">1</span><span class="p">)):</span>
- <span class="n">conf_mat</span><span class="p">[</span><span class="n">ytrue</span><span class="o">.</span><span class="n">long</span><span class="p">(),</span> <span class="n">ypred</span><span class="o">.</span><span class="n">long</span><span class="p">()]</span> <span class="o">+=</span> <span class="mi">1</span>
-
- <span class="c1"># reconstruct the entire scene</span>
- <span class="n">inputs</span> <span class="o">=</span> <span class="n">reconstruct_scene</span><span class="p">(</span><span class="n">inp</span><span class="p">)</span>
- <span class="n">labels</span> <span class="o">=</span> <span class="n">reconstruct_scene</span><span class="p">(</span><span class="n">lab</span><span class="p">)</span>
- <span class="n">prdtcn</span> <span class="o">=</span> <span class="n">reconstruct_scene</span><span class="p">(</span><span class="n">prd</span><span class="p">)</span>
-
- <span class="c1"># print progress</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Scene </span><span class="si">{:d}</span><span class="s1">/</span><span class="si">{:d}</span><span class="s1">, Id: </span><span class="si">{}</span><span class="s1">, Accuracy: </span><span class="si">{:.2f}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span>
- <span class="n">i</span> <span class="o">+</span> <span class="mi">1</span><span class="p">,</span> <span class="nb">len</span><span class="p">(</span><span class="n">scene_ids</span><span class="p">),</span> <span class="n">sid</span><span class="p">,</span> <span class="n">accuracy_function</span><span class="p">(</span><span class="n">prdtcn</span><span class="p">,</span> <span class="n">labels</span><span class="p">)))</span>
-
- <span class="c1"># save outputs to dictionary</span>
- <span class="n">output</span><span class="p">[</span><span class="n">sid</span><span class="p">]</span> <span class="o">=</span> <span class="p">{</span><span class="s1">'input'</span><span class="p">:</span> <span class="n">inputs</span><span class="p">,</span> <span class="s1">'labels'</span><span class="p">:</span> <span class="n">labels</span><span class="p">,</span> <span class="s1">'prediction'</span><span class="p">:</span> <span class="n">prdtcn</span><span class="p">}</span>
-
- <span class="c1"># plot current scene</span>
- <span class="k">if</span> <span class="n">plot</span><span class="p">:</span>
- <span class="n">fig</span><span class="p">,</span> <span class="n">ax</span> <span class="o">=</span> <span class="n">plot_sample</span><span class="p">(</span><span class="n">inputs</span><span class="o">.</span><span class="n">clip</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="mi">1</span><span class="p">),</span>
- <span class="n">ds</span><span class="o">.</span><span class="n">dataset</span><span class="o">.</span><span class="n">use_bands</span><span class="p">,</span>
- <span class="n">ds</span><span class="o">.</span><span class="n">dataset</span><span class="o">.</span><span class="n">labels</span><span class="p">,</span>
- <span class="n">y</span><span class="o">=</span><span class="n">labels</span><span class="p">,</span>
- <span class="n">y_pred</span><span class="o">=</span><span class="n">prdtcn</span><span class="p">,</span>
- <span class="n">state</span><span class="o">=</span><span class="n">sname</span><span class="p">,</span>
- <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span>
-
- <span class="k">return</span> <span class="n">output</span><span class="p">,</span> <span class="n">conf_mat</span></div>
-</pre></div>
-
- </div>
-
- </div>
- <footer>
-
-
- <hr/>
-
- <div role="contentinfo">
- <p>
-
- © Copyright 2020, Daniel Frisinghelli
-
- </p>
- </div>
-
-
-
- Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a
-
- <a href="https://github.com/rtfd/sphinx_rtd_theme">theme</a>
-
- provided by <a href="https://readthedocs.org">Read the Docs</a>.
-
-</footer>
-
- </div>
- </div>
-
- </section>
-
- </div>
-
-
- <script type="text/javascript">
- jQuery(function () {
- SphinxRtdTheme.Navigation.enable(true);
- });
- </script>
-
-
-
-
-
-
-</body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/_modules/core/split.html b/docs/_build/html/_modules/core/split.html
deleted file mode 100644
index 452103871c0eb76a67777b8c3bd3bfb5434709e1..0000000000000000000000000000000000000000
--- a/docs/_build/html/_modules/core/split.html
+++ /dev/null
@@ -1,885 +0,0 @@
-
-
-<!DOCTYPE html>
-<html class="writer-html5" lang="en" >
-<head>
- <meta charset="utf-8">
-
- <meta name="viewport" content="width=device-width, initial-scale=1.0">
-
- <title>core.split — PySegCNN documentation</title>
-
-
-
- <link rel="stylesheet" href="../../_static/css/theme.css" type="text/css" />
- <link rel="stylesheet" href="../../_static/pygments.css" type="text/css" />
-
-
-
-
-
-
-
- <!--[if lt IE 9]>
- <script src="../../_static/js/html5shiv.min.js"></script>
- <![endif]-->
-
-
- <script type="text/javascript" id="documentation_options" data-url_root="../../" src="../../_static/documentation_options.js"></script>
- <script src="../../_static/jquery.js"></script>
- <script src="../../_static/underscore.js"></script>
- <script src="../../_static/doctools.js"></script>
- <script src="../../_static/language_data.js"></script>
-
- <script type="text/javascript" src="../../_static/js/theme.js"></script>
-
-
- <link rel="index" title="Index" href="../../genindex.html" />
- <link rel="search" title="Search" href="../../search.html" />
-</head>
-
-<body class="wy-body-for-nav">
-
-
- <div class="wy-grid-for-nav">
-
- <nav data-toggle="wy-nav-shift" class="wy-nav-side">
- <div class="wy-side-scroll">
- <div class="wy-side-nav-search" >
-
-
-
- <a href="../../index.html" class="icon icon-home" alt="Documentation Home"> PySegCNN
-
-
-
- </a>
-
-
-
-
-
-
-
-<div role="search">
- <form id="rtd-search-form" class="wy-form" action="../../search.html" method="get">
- <input type="text" name="q" placeholder="Search docs" />
- <input type="hidden" name="check_keywords" value="yes" />
- <input type="hidden" name="area" value="default" />
- </form>
-</div>
-
-
- </div>
-
-
- <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
-
-
-
-
-
-
- <ul>
-<li class="toctree-l1"><a class="reference internal" href="../../source/requirements.html">Requirements</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../../source/installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../../source/api.html">API Reference</a></li>
-</ul>
-
-
-
- </div>
-
- </div>
- </nav>
-
- <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
-
-
- <nav class="wy-nav-top" aria-label="top navigation">
-
- <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
- <a href="../../index.html">PySegCNN</a>
-
- </nav>
-
-
- <div class="wy-nav-content">
-
- <div class="rst-content">
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-<div role="navigation" aria-label="breadcrumbs navigation">
-
- <ul class="wy-breadcrumbs">
-
- <li><a href="../../index.html" class="icon icon-home"></a> »</li>
-
- <li><a href="../index.html">Module code</a> »</li>
-
- <li>core.split</li>
-
-
- <li class="wy-breadcrumbs-aside">
-
- </li>
-
- </ul>
-
-
- <hr/>
-</div>
- <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
- <div itemprop="articleBody">
-
- <h1>Source code for core.split</h1><div class="highlight"><pre>
-<span></span><span class="sd">"""Split the dataset into training, validation and test set.</span>
-
-<span class="sd">License</span>
-<span class="sd">-------</span>
-
-<span class="sd"> Copyright (c) 2020 Daniel Frisinghelli</span>
-
-<span class="sd"> This source code is licensed under the GNU General Public License v3.</span>
-
-<span class="sd"> See the LICENSE file in the repository's root directory.</span>
-
-<span class="sd">"""</span>
-
-<span class="c1"># !/usr/bin/env python</span>
-<span class="c1"># -*- coding: utf-8 -*-</span>
-
-<span class="c1"># builtins</span>
-<span class="kn">import</span> <span class="nn">datetime</span>
-<span class="kn">import</span> <span class="nn">enum</span>
-
-<span class="c1"># externals</span>
-<span class="kn">import</span> <span class="nn">numpy</span> <span class="k">as</span> <span class="nn">np</span>
-<span class="kn">from</span> <span class="nn">torch.utils.data.dataset</span> <span class="kn">import</span> <span class="n">Subset</span>
-
-<span class="c1"># the names of the subsets</span>
-<span class="n">SUBSET_NAMES</span> <span class="o">=</span> <span class="p">[</span><span class="s1">'train'</span><span class="p">,</span> <span class="s1">'valid'</span><span class="p">,</span> <span class="s1">'test'</span><span class="p">]</span>
-
-
-<span class="k">def</span> <span class="nf">_ds_len</span><span class="p">(</span><span class="n">ds</span><span class="p">,</span> <span class="n">ratio</span><span class="p">):</span>
- <span class="sd">"""Calcute number of samples in a dataset given a ratio.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> ds : `collections.Sized`</span>
-<span class="sd"> An object with a __len__ attribute.</span>
-<span class="sd"> ratio : `float`</span>
-<span class="sd"> A ratio to multiply with the length of ``ds``.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> n_samples: `int`</span>
-<span class="sd"> Length of ``ds`` * ``ratio``.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="nb">int</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">round</span><span class="p">(</span><span class="nb">len</span><span class="p">(</span><span class="n">ds</span><span class="p">)</span> <span class="o">*</span> <span class="n">ratio</span><span class="p">))</span>
-
-
-<div class="viewcode-block" id="random_tile_split"><a class="viewcode-back" href="../../source/core.html#core.split.random_tile_split">[docs]</a><span class="k">def</span> <span class="nf">random_tile_split</span><span class="p">(</span><span class="n">ds</span><span class="p">,</span> <span class="n">tvratio</span><span class="p">,</span> <span class="n">ttratio</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">0</span><span class="p">):</span>
- <span class="sd">"""Randomly split the tiles of a dataset.</span>
-
-<span class="sd"> For each scene, the tiles of the scene can be distributed among the</span>
-<span class="sd"> training, validation and test set.</span>
-
-<span class="sd"> The parameters ``ttratio`` and ``tvratio`` control the size of the</span>
-<span class="sd"> training, validation and test datasets.</span>
-
-<span class="sd"> Test dataset size : (1 - ``ttratio``) * len(``ds``)</span>
-<span class="sd"> Train dataset size : ``ttratio`` * ``tvratio`` * len(``ds``)</span>
-<span class="sd"> Validation dataset size: ``ttratio`` * (1 - ``tvratio``) * len(``ds``)</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> ds : `pysegcnn.core.dataset.ImageDataset`</span>
-<span class="sd"> An instance of `~pysegcnn.core.dataset.ImageDataset`.</span>
-<span class="sd"> tvratio : `float`</span>
-<span class="sd"> The ratio of training data to validation data, e.g. ``tvratio`` = 0.8</span>
-<span class="sd"> means 80% training, 20% validation.</span>
-<span class="sd"> ttratio : `float`, optional</span>
-<span class="sd"> The ratio of training and validation data to test data, e.g.</span>
-<span class="sd"> ``ttratio`` = 0.6 means 60% for training and validation, 40% for</span>
-<span class="sd"> testing. The default is 1.</span>
-<span class="sd"> seed : `int`, optional</span>
-<span class="sd"> The random seed for reproducibility. The default is 0.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> AssertionError</span>
-<span class="sd"> Raised if the splits are not pairwise disjoint.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> subsets : `dict`</span>
-<span class="sd"> Subset dictionary with keys:</span>
-<span class="sd"> ``'train'``</span>
-<span class="sd"> dictionary containing the training scenes.</span>
-<span class="sd"> ``'valid'``</span>
-<span class="sd"> dictionary containing the validation scenes.</span>
-<span class="sd"> ``'test'``</span>
-<span class="sd"> dictionary containing the test scenes.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># set the random seed for reproducibility</span>
- <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">seed</span><span class="p">(</span><span class="n">seed</span><span class="p">)</span>
-
- <span class="c1"># randomly permute indices to access dataset</span>
- <span class="n">indices</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">permutation</span><span class="p">(</span><span class="nb">len</span><span class="p">(</span><span class="n">ds</span><span class="p">))</span>
-
- <span class="c1"># length of the training and validation dataset</span>
- <span class="c1"># number of samples: (ttratio * len(ds))</span>
- <span class="n">trav_len</span> <span class="o">=</span> <span class="n">_ds_len</span><span class="p">(</span><span class="n">indices</span><span class="p">,</span> <span class="n">ttratio</span><span class="p">)</span>
- <span class="n">trav_indices</span> <span class="o">=</span> <span class="n">indices</span><span class="p">[:</span><span class="n">trav_len</span><span class="p">]</span>
-
- <span class="c1"># length of the training dataset</span>
- <span class="c1"># number of samples: (ttratio * tvratio * len(ds))</span>
- <span class="n">train_len</span> <span class="o">=</span> <span class="n">_ds_len</span><span class="p">(</span><span class="n">trav_indices</span><span class="p">,</span> <span class="n">tvratio</span><span class="p">)</span>
- <span class="n">train_ind</span> <span class="o">=</span> <span class="n">trav_indices</span><span class="p">[:</span><span class="n">train_len</span><span class="p">]</span>
-
- <span class="c1"># length of the validation dataset</span>
- <span class="c1"># number of samples: (ttratio * (1- tvratio) * len(ds))</span>
- <span class="n">valid_ind</span> <span class="o">=</span> <span class="n">trav_indices</span><span class="p">[</span><span class="n">train_len</span><span class="p">:]</span>
-
- <span class="c1"># length of the test dataset</span>
- <span class="c1"># number of samples: ((1 - ttratio) * len(ds))</span>
- <span class="n">test_ind</span> <span class="o">=</span> <span class="n">indices</span><span class="p">[</span><span class="n">trav_len</span><span class="p">:]</span>
-
- <span class="c1"># get the tiles of the scenes of each dataset</span>
- <span class="n">subsets</span> <span class="o">=</span> <span class="p">{}</span>
- <span class="k">for</span> <span class="n">name</span><span class="p">,</span> <span class="n">dataset</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">([</span><span class="n">train_ind</span><span class="p">,</span> <span class="n">valid_ind</span><span class="p">,</span> <span class="n">test_ind</span><span class="p">]):</span>
-
- <span class="c1"># store the indices and corresponding tiles of the current subset to</span>
- <span class="c1"># dictionary</span>
- <span class="n">subsets</span><span class="p">[</span><span class="n">SUBSET_NAMES</span><span class="p">[</span><span class="n">name</span><span class="p">]]</span> <span class="o">=</span> <span class="p">{</span><span class="n">k</span><span class="p">:</span> <span class="n">ds</span><span class="o">.</span><span class="n">scenes</span><span class="p">[</span><span class="n">k</span><span class="p">]</span> <span class="k">for</span> <span class="n">k</span> <span class="ow">in</span> <span class="n">dataset</span><span class="p">}</span>
-
- <span class="c1"># check if the splits are disjoint</span>
- <span class="k">assert</span> <span class="n">pairwise_disjoint</span><span class="p">([</span><span class="n">s</span><span class="o">.</span><span class="n">keys</span><span class="p">()</span> <span class="k">for</span> <span class="n">s</span> <span class="ow">in</span> <span class="n">subsets</span><span class="o">.</span><span class="n">values</span><span class="p">()])</span>
-
- <span class="k">return</span> <span class="n">subsets</span></div>
-
-
-<div class="viewcode-block" id="random_scene_split"><a class="viewcode-back" href="../../source/core.html#core.split.random_scene_split">[docs]</a><span class="k">def</span> <span class="nf">random_scene_split</span><span class="p">(</span><span class="n">ds</span><span class="p">,</span> <span class="n">tvratio</span><span class="p">,</span> <span class="n">ttratio</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">0</span><span class="p">):</span>
- <span class="sd">"""Randomly split the tiles of a dataset.</span>
-
-<span class="sd"> For each scene, all the tiles of the scene are included in either the</span>
-<span class="sd"> training, validation or test set, respectively.</span>
-
-<span class="sd"> The parameters ``ttratio`` and ``tvratio`` control the size of the</span>
-<span class="sd"> training, validation and test datasets.</span>
-
-<span class="sd"> Test dataset size : (1 - ``ttratio``) * len(``ds``)</span>
-<span class="sd"> Train dataset size : ``ttratio`` * ``tvratio`` * len(``ds``)</span>
-<span class="sd"> Validation dataset size: ``ttratio`` * (1 - ``tvratio``) * len(``ds``)</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> ds : `pysegcnn.core.dataset.ImageDataset`</span>
-<span class="sd"> An instance of `~pysegcnn.core.dataset.ImageDataset`.</span>
-<span class="sd"> tvratio : `float`</span>
-<span class="sd"> The ratio of training data to validation data, e.g. ``tvratio`` = 0.8</span>
-<span class="sd"> means 80% training, 20% validation.</span>
-<span class="sd"> ttratio : `float`, optional</span>
-<span class="sd"> The ratio of training and validation data to test data, e.g.</span>
-<span class="sd"> ``ttratio`` = 0.6 means 60% for training and validation, 40% for</span>
-<span class="sd"> testing. The default is 1.</span>
-<span class="sd"> seed : `int`, optional</span>
-<span class="sd"> The random seed for reproducibility. The default is 0.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> AssertionError</span>
-<span class="sd"> Raised if the splits are not pairwise disjoint.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> subsets : `dict`</span>
-<span class="sd"> Subset dictionary with keys:</span>
-<span class="sd"> ``'train'``</span>
-<span class="sd"> dictionary containing the training scenes.</span>
-<span class="sd"> ``'valid'``</span>
-<span class="sd"> dictionary containing the validation scenes.</span>
-<span class="sd"> ``'test'``</span>
-<span class="sd"> dictionary containing the test scenes.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># set the random seed for reproducibility</span>
- <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">seed</span><span class="p">(</span><span class="n">seed</span><span class="p">)</span>
-
- <span class="c1"># get the names of the scenes and generate random permutation</span>
- <span class="n">scene_ids</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">permutation</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">unique</span><span class="p">([</span><span class="n">s</span><span class="p">[</span><span class="s1">'id'</span><span class="p">]</span> <span class="k">for</span> <span class="n">s</span> <span class="ow">in</span> <span class="n">ds</span><span class="o">.</span><span class="n">scenes</span><span class="p">]))</span>
-
- <span class="c1"># the training and validation scenes</span>
- <span class="c1"># number of samples: (ttratio * nscenes)</span>
- <span class="n">trav_len</span> <span class="o">=</span> <span class="n">_ds_len</span><span class="p">(</span><span class="n">scene_ids</span><span class="p">,</span> <span class="n">ttratio</span><span class="p">)</span>
- <span class="n">trav_scenes</span> <span class="o">=</span> <span class="n">scene_ids</span><span class="p">[:</span><span class="n">trav_len</span><span class="p">]</span>
-
- <span class="c1"># the training scenes</span>
- <span class="c1"># number of samples: (ttratio * tvratio * nscenes)</span>
- <span class="n">train_len</span> <span class="o">=</span> <span class="n">_ds_len</span><span class="p">(</span><span class="n">trav_scenes</span><span class="p">,</span> <span class="n">tvratio</span><span class="p">)</span>
- <span class="n">train_scenes</span> <span class="o">=</span> <span class="n">trav_scenes</span><span class="p">[:</span><span class="n">train_len</span><span class="p">]</span>
-
- <span class="c1"># the validation scenes</span>
- <span class="c1"># number of samples: (ttratio * (1- tvratio) * nscenes)</span>
- <span class="n">valid_scenes</span> <span class="o">=</span> <span class="n">trav_scenes</span><span class="p">[</span><span class="n">train_len</span><span class="p">:]</span>
-
- <span class="c1"># the test scenes</span>
- <span class="c1"># number of samples:((1 - ttratio) * nscenes)</span>
- <span class="n">test_scenes</span> <span class="o">=</span> <span class="n">scene_ids</span><span class="p">[</span><span class="n">trav_len</span><span class="p">:]</span>
-
- <span class="c1"># get the tiles of the scenes of each dataset</span>
- <span class="n">subsets</span> <span class="o">=</span> <span class="p">{}</span>
- <span class="k">for</span> <span class="n">name</span><span class="p">,</span> <span class="n">dataset</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">([</span><span class="n">train_scenes</span><span class="p">,</span> <span class="n">valid_scenes</span><span class="p">,</span> <span class="n">test_scenes</span><span class="p">]):</span>
-
- <span class="c1"># store the indices and corresponding tiles of the current subset to</span>
- <span class="c1"># dictionary</span>
- <span class="n">subsets</span><span class="p">[</span><span class="n">SUBSET_NAMES</span><span class="p">[</span><span class="n">name</span><span class="p">]]</span> <span class="o">=</span> <span class="p">{</span><span class="n">k</span><span class="p">:</span> <span class="n">v</span> <span class="k">for</span> <span class="n">k</span><span class="p">,</span> <span class="n">v</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">(</span><span class="n">ds</span><span class="o">.</span><span class="n">scenes</span><span class="p">)</span>
- <span class="k">if</span> <span class="n">v</span><span class="p">[</span><span class="s1">'id'</span><span class="p">]</span> <span class="ow">in</span> <span class="n">dataset</span><span class="p">}</span>
-
- <span class="c1"># check if the splits are disjoint</span>
- <span class="k">assert</span> <span class="n">pairwise_disjoint</span><span class="p">([</span><span class="n">s</span><span class="o">.</span><span class="n">keys</span><span class="p">()</span> <span class="k">for</span> <span class="n">s</span> <span class="ow">in</span> <span class="n">subsets</span><span class="o">.</span><span class="n">values</span><span class="p">()])</span>
-
- <span class="k">return</span> <span class="n">subsets</span></div>
-
-
-<span class="c1"># split the scenes of a dataset based on a date, useful for time series data</span>
-<span class="c1"># scenes before date build the training set, scenes after the date build the</span>
-<span class="c1"># validation set, the test set is empty</span>
-<div class="viewcode-block" id="date_scene_split"><a class="viewcode-back" href="../../source/core.html#core.split.date_scene_split">[docs]</a><span class="k">def</span> <span class="nf">date_scene_split</span><span class="p">(</span><span class="n">ds</span><span class="p">,</span> <span class="n">date</span><span class="p">,</span> <span class="n">dateformat</span><span class="o">=</span><span class="s1">'%Y%m</span><span class="si">%d</span><span class="s1">'</span><span class="p">):</span>
- <span class="sd">"""Split the dataset based on a date.</span>
-
-<span class="sd"> Scenes before ``date`` build the training set, scenes after ``date`` build</span>
-<span class="sd"> the validation set, the test set is empty.</span>
-
-<span class="sd"> Useful for time series data.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> ds : `pysegcnn.core.dataset.ImageDataset`</span>
-<span class="sd"> An instance of `~pysegcnn.core.dataset.ImageDataset`.</span>
-<span class="sd"> date : `str`</span>
-<span class="sd"> A date.</span>
-<span class="sd"> dateformat : `str`, optional</span>
-<span class="sd"> The format of ``date``. ``dateformat`` is used by</span>
-<span class="sd"> `datetime.datetime.strptime' to parse ``date`` to a `datetime.datetime`</span>
-<span class="sd"> object. The default is '%Y%m%d'.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> AssertionError</span>
-<span class="sd"> Raised if the splits are not pairwise disjoint.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> subsets : `dict`</span>
-<span class="sd"> Subset dictionary with keys:</span>
-<span class="sd"> ``'train'``</span>
-<span class="sd"> dictionary containing the training scenes.</span>
-<span class="sd"> ``'valid'``</span>
-<span class="sd"> dictionary containing the validation scenes.</span>
-<span class="sd"> ``'test'``</span>
-<span class="sd"> dictionary containing the test scenes, empty.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># convert date to datetime object</span>
- <span class="n">date</span> <span class="o">=</span> <span class="n">datetime</span><span class="o">.</span><span class="n">datetime</span><span class="o">.</span><span class="n">strptime</span><span class="p">(</span><span class="n">date</span><span class="p">,</span> <span class="n">dateformat</span><span class="p">)</span>
-
- <span class="c1"># the training, validation and test scenes</span>
- <span class="n">train_scenes</span> <span class="o">=</span> <span class="p">{</span><span class="n">i</span><span class="p">:</span> <span class="n">s</span> <span class="k">for</span> <span class="n">i</span><span class="p">,</span> <span class="n">s</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">(</span><span class="n">ds</span><span class="o">.</span><span class="n">scenes</span><span class="p">)</span> <span class="k">if</span> <span class="n">s</span><span class="p">[</span><span class="s1">'date'</span><span class="p">]</span> <span class="o"><=</span> <span class="n">date</span><span class="p">}</span>
- <span class="n">valid_scenes</span> <span class="o">=</span> <span class="p">{</span><span class="n">i</span><span class="p">:</span> <span class="n">s</span> <span class="k">for</span> <span class="n">i</span><span class="p">,</span> <span class="n">s</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">(</span><span class="n">ds</span><span class="o">.</span><span class="n">scenes</span><span class="p">)</span> <span class="k">if</span> <span class="n">s</span><span class="p">[</span><span class="s1">'date'</span><span class="p">]</span> <span class="o">></span> <span class="n">date</span><span class="p">}</span>
- <span class="n">test_scenes</span> <span class="o">=</span> <span class="p">{}</span>
-
- <span class="c1"># build the training and test datasets</span>
- <span class="n">subsets</span> <span class="o">=</span> <span class="p">{}</span>
- <span class="k">for</span> <span class="n">name</span><span class="p">,</span> <span class="n">scenes</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">([</span><span class="n">train_scenes</span><span class="p">,</span> <span class="n">valid_scenes</span><span class="p">,</span> <span class="n">test_scenes</span><span class="p">]):</span>
-
- <span class="c1"># store the indices and corresponding tiles of the current subset to</span>
- <span class="c1"># dictionary</span>
- <span class="n">subsets</span><span class="p">[</span><span class="n">SUBSET_NAMES</span><span class="p">[</span><span class="n">name</span><span class="p">]]</span> <span class="o">=</span> <span class="n">scenes</span>
-
- <span class="c1"># check if the splits are disjoint</span>
- <span class="k">assert</span> <span class="n">pairwise_disjoint</span><span class="p">([</span><span class="n">s</span><span class="o">.</span><span class="n">keys</span><span class="p">()</span> <span class="k">for</span> <span class="n">s</span> <span class="ow">in</span> <span class="n">subsets</span><span class="o">.</span><span class="n">values</span><span class="p">()])</span>
-
- <span class="k">return</span> <span class="n">subsets</span></div>
-
-
-<div class="viewcode-block" id="pairwise_disjoint"><a class="viewcode-back" href="../../source/core.html#core.split.pairwise_disjoint">[docs]</a><span class="k">def</span> <span class="nf">pairwise_disjoint</span><span class="p">(</span><span class="n">sets</span><span class="p">):</span>
- <span class="sd">"""Check if ``sets`` are pairwise disjoint.</span>
-
-<span class="sd"> Sets are pairwise disjoint if the length of their union equals the sum of</span>
-<span class="sd"> their lengths.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> sets : `list` [`collections.Sized`]</span>
-<span class="sd"> A list of sized objects.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> disjoint : `bool`</span>
-<span class="sd"> Whether the sets are pairwise disjoint.</span>
-
-<span class="sd"> """</span>
- <span class="n">union</span> <span class="o">=</span> <span class="nb">set</span><span class="p">()</span><span class="o">.</span><span class="n">union</span><span class="p">(</span><span class="o">*</span><span class="n">sets</span><span class="p">)</span>
- <span class="n">n</span> <span class="o">=</span> <span class="nb">sum</span><span class="p">(</span><span class="nb">len</span><span class="p">(</span><span class="n">u</span><span class="p">)</span> <span class="k">for</span> <span class="n">u</span> <span class="ow">in</span> <span class="n">sets</span><span class="p">)</span>
- <span class="k">return</span> <span class="n">n</span> <span class="o">==</span> <span class="nb">len</span><span class="p">(</span><span class="n">union</span><span class="p">)</span></div>
-
-
-<div class="viewcode-block" id="CustomSubset"><a class="viewcode-back" href="../../source/core.html#core.split.CustomSubset">[docs]</a><span class="k">class</span> <span class="nc">CustomSubset</span><span class="p">(</span><span class="n">Subset</span><span class="p">):</span>
- <span class="sd">"""Custom subset inheriting `torch.utils.data.Subset`."""</span>
-
- <span class="k">def</span> <span class="fm">__repr__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Representation of ``~pysegcnn.core.split.CustomSubset``."""</span>
- <span class="c1"># representation string</span>
- <span class="n">fs</span> <span class="o">=</span> <span class="s1">'- </span><span class="si">{}</span><span class="s1">: </span><span class="si">{:d}</span><span class="s1"> tiles (</span><span class="si">{:.2f}</span><span class="s1">%)'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">name</span><span class="p">,</span> <span class="nb">len</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">scenes</span><span class="p">),</span> <span class="mi">100</span> <span class="o">*</span> <span class="nb">len</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">scenes</span><span class="p">)</span> <span class="o">/</span>
- <span class="nb">len</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">dataset</span><span class="p">))</span>
-
- <span class="k">return</span> <span class="n">fs</span></div>
-
-
-<div class="viewcode-block" id="SceneSubset"><a class="viewcode-back" href="../../source/core.html#core.split.SceneSubset">[docs]</a><span class="k">class</span> <span class="nc">SceneSubset</span><span class="p">(</span><span class="n">CustomSubset</span><span class="p">):</span>
- <span class="sd">"""A custom subset for dataset splits where the scenes are preserved.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> ds : `pysegcnn.core.dataset.ImageDataset`</span>
-<span class="sd"> An instance of `~pysegcnn.core.dataset.ImageDataset`.</span>
-<span class="sd"> indices : `list` [`int`]</span>
-<span class="sd"> List of the subset indices to access ``ds``.</span>
-<span class="sd"> name : `str`</span>
-<span class="sd"> Name of the subset.</span>
-<span class="sd"> scenes : `list` [`dict`]</span>
-<span class="sd"> List of the subset tiles.</span>
-<span class="sd"> scene_ids : `list` or `numpy.ndarray`</span>
-<span class="sd"> Container of the scene ids.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">ds</span><span class="p">,</span> <span class="n">indices</span><span class="p">,</span> <span class="n">name</span><span class="p">,</span> <span class="n">scenes</span><span class="p">,</span> <span class="n">scene_ids</span><span class="p">):</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">dataset</span><span class="o">=</span><span class="n">ds</span><span class="p">,</span> <span class="n">indices</span><span class="o">=</span><span class="n">indices</span><span class="p">)</span>
-
- <span class="c1"># the name of the subset</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">name</span> <span class="o">=</span> <span class="n">name</span>
-
- <span class="c1"># the scene in the subset</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">scenes</span> <span class="o">=</span> <span class="n">scenes</span>
-
- <span class="c1"># the names of the scenes</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">ids</span> <span class="o">=</span> <span class="n">scene_ids</span></div>
-
-
-<div class="viewcode-block" id="RandomSubset"><a class="viewcode-back" href="../../source/core.html#core.split.RandomSubset">[docs]</a><span class="k">class</span> <span class="nc">RandomSubset</span><span class="p">(</span><span class="n">CustomSubset</span><span class="p">):</span>
- <span class="sd">"""A custom subset for random dataset splits.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> ds : `pysegcnn.core.dataset.ImageDataset`</span>
-<span class="sd"> An instance of `~pysegcnn.core.dataset.ImageDataset`.</span>
-<span class="sd"> indices : `list` [`int`]</span>
-<span class="sd"> List of the subset indices to access ``ds``.</span>
-<span class="sd"> name : `str`</span>
-<span class="sd"> Name of the subset.</span>
-<span class="sd"> scenes : `list` [`dict`]</span>
-<span class="sd"> List of the subset tiles.</span>
-<span class="sd"> scene_ids : `list` or `numpy.ndarray`</span>
-<span class="sd"> Container of the scene ids.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">ds</span><span class="p">,</span> <span class="n">indices</span><span class="p">,</span> <span class="n">name</span><span class="p">,</span> <span class="n">scenes</span><span class="p">,</span> <span class="n">scene_ids</span><span class="p">):</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">dataset</span><span class="o">=</span><span class="n">ds</span><span class="p">,</span> <span class="n">indices</span><span class="o">=</span><span class="n">indices</span><span class="p">)</span>
-
- <span class="c1"># the name of the subset</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">name</span> <span class="o">=</span> <span class="n">name</span>
-
- <span class="c1"># the scene in the subset</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">scenes</span> <span class="o">=</span> <span class="n">scenes</span></div>
-
-
-<div class="viewcode-block" id="Split"><a class="viewcode-back" href="../../source/core.html#core.split.Split">[docs]</a><span class="k">class</span> <span class="nc">Split</span><span class="p">(</span><span class="nb">object</span><span class="p">):</span>
- <span class="sd">"""Generic class handling how ``ds`` is split.</span>
-
-<span class="sd"> Inherit `~pysegcnn.core.split.Split` and implement the</span>
-<span class="sd"> `~pysegcnn.core.split.Split.subsets` and</span>
-<span class="sd"> `~pysegcnn.core.split.Split.subset_type` method.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> ds : `pysegcnn.core.dataset.ImageDataset`</span>
-<span class="sd"> An instance of `~pysegcnn.core.dataset.ImageDataset`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">ds</span><span class="p">):</span>
-
- <span class="c1"># the dataset to split</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">ds</span> <span class="o">=</span> <span class="n">ds</span>
-
-<div class="viewcode-block" id="Split.split"><a class="viewcode-back" href="../../source/core.html#core.split.Split.split">[docs]</a> <span class="k">def</span> <span class="nf">split</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Split dataset into training, validation and test set.</span>
-
-<span class="sd"> `~pysegcnn.core.split.Split.split` works only if</span>
-<span class="sd"> `~pysegcnn.core.split.Split.subsets` and</span>
-<span class="sd"> `~pysegcnn.core.split.Split.subset_type` are implemented.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># build the subsets</span>
- <span class="n">ds_split</span> <span class="o">=</span> <span class="p">[]</span>
- <span class="k">for</span> <span class="n">name</span><span class="p">,</span> <span class="n">sub</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">subsets</span><span class="p">()</span><span class="o">.</span><span class="n">items</span><span class="p">():</span>
-
- <span class="c1"># the scene identifiers of the current subset</span>
- <span class="n">ids</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">unique</span><span class="p">([</span><span class="n">s</span><span class="p">[</span><span class="s1">'id'</span><span class="p">]</span> <span class="k">for</span> <span class="n">s</span> <span class="ow">in</span> <span class="n">sub</span><span class="o">.</span><span class="n">values</span><span class="p">()])</span>
-
- <span class="c1"># build the subset</span>
- <span class="n">sbst</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">subset_type</span><span class="p">()(</span><span class="bp">self</span><span class="o">.</span><span class="n">ds</span><span class="p">,</span> <span class="nb">list</span><span class="p">(</span><span class="n">sub</span><span class="o">.</span><span class="n">keys</span><span class="p">()),</span> <span class="n">name</span><span class="p">,</span>
- <span class="nb">list</span><span class="p">(</span><span class="n">sub</span><span class="o">.</span><span class="n">values</span><span class="p">()),</span> <span class="n">ids</span><span class="p">)</span>
- <span class="n">ds_split</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">sbst</span><span class="p">)</span>
-
- <span class="k">return</span> <span class="n">ds_split</span></div>
-
-<div class="viewcode-block" id="Split.subsets"><a class="viewcode-back" href="../../source/core.html#core.split.Split.subsets">[docs]</a> <span class="k">def</span> <span class="nf">subsets</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Define training, validation and test sets.</span>
-
-<span class="sd"> Wrapper method for</span>
-<span class="sd"> `pysegcnn.core.split.Split.random_tile_split`,</span>
-<span class="sd"> `pysegcnn.core.split.Split.random_scene_split` or</span>
-<span class="sd"> `pysegcnn.core.split.Split.date_scene_split`.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> NotImplementedError</span>
-<span class="sd"> Raised if `pysegcnn.core.split.Split` is not inherited.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
- <span class="k">raise</span> <span class="ne">NotImplementedError</span></div>
-
-<div class="viewcode-block" id="Split.subset_type"><a class="viewcode-back" href="../../source/core.html#core.split.Split.subset_type">[docs]</a> <span class="k">def</span> <span class="nf">subset_type</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Define the type of each subset.</span>
-
-<span class="sd"> Wrapper method for</span>
-<span class="sd"> `pysegcnn.core.split.RandomSubset` or</span>
-<span class="sd"> `pysegcnn.core.split.SceneSubset`.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> NotImplementedError</span>
-<span class="sd"> Raised if `pysegcnn.core.split.Split` is not inherited.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
- <span class="k">raise</span> <span class="ne">NotImplementedError</span></div></div>
-
-
-<div class="viewcode-block" id="DateSplit"><a class="viewcode-back" href="../../source/core.html#core.split.DateSplit">[docs]</a><span class="k">class</span> <span class="nc">DateSplit</span><span class="p">(</span><span class="n">Split</span><span class="p">):</span>
- <span class="sd">"""Split the dataset based on a date.</span>
-
-<span class="sd"> Class wrapper for `pysegcnn.core.split.Split.date_scene_split`.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> ds : `pysegcnn.core.dataset.ImageDataset`</span>
-<span class="sd"> An instance of `~pysegcnn.core.dataset.ImageDataset`.</span>
-<span class="sd"> date : 'str'</span>
-<span class="sd"> A date.</span>
-<span class="sd"> dateformat : 'str', optional</span>
-<span class="sd"> The format of ``date``. ``dateformat`` is used by</span>
-<span class="sd"> `datetime.datetime.strptime' to parse ``date`` to a `datetime.datetime`</span>
-<span class="sd"> object. The default is '%Y%m%d'.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">ds</span><span class="p">,</span> <span class="n">date</span><span class="p">,</span> <span class="n">dateformat</span><span class="p">):</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">ds</span><span class="p">)</span>
-
- <span class="c1"># the date to split the dataset</span>
- <span class="c1"># before: training set</span>
- <span class="c1"># after : validation set</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">date</span> <span class="o">=</span> <span class="n">date</span>
-
- <span class="c1"># the format of the date</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">dateformat</span> <span class="o">=</span> <span class="n">dateformat</span>
-
-<div class="viewcode-block" id="DateSplit.subsets"><a class="viewcode-back" href="../../source/core.html#core.split.DateSplit.subsets">[docs]</a> <span class="k">def</span> <span class="nf">subsets</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Wrap `pysegcnn.core.split.Split.date_scene_split`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> subsets : `dict`</span>
-<span class="sd"> Subset dictionary with keys:</span>
-<span class="sd"> ``'train'``</span>
-<span class="sd"> dictionary containing the training scenes.</span>
-<span class="sd"> ``'valid'``</span>
-<span class="sd"> dictionary containing the validation scenes.</span>
-<span class="sd"> ``'test'``</span>
-<span class="sd"> dictionary containing the test scenes, empty.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="n">date_scene_split</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">ds</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">date</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">dateformat</span><span class="p">)</span></div>
-
-<div class="viewcode-block" id="DateSplit.subset_type"><a class="viewcode-back" href="../../source/core.html#core.split.DateSplit.subset_type">[docs]</a> <span class="k">def</span> <span class="nf">subset_type</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Wrap `pysegcnn.core.split.SceneSubset`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> SceneSubset : `pysegcnn.core.split.SceneSubset`</span>
-<span class="sd"> The subset type.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="n">SceneSubset</span></div></div>
-
-
-<div class="viewcode-block" id="RandomSplit"><a class="viewcode-back" href="../../source/core.html#core.split.RandomSplit">[docs]</a><span class="k">class</span> <span class="nc">RandomSplit</span><span class="p">(</span><span class="n">Split</span><span class="p">):</span>
- <span class="sd">"""Randomly split the dataset.</span>
-
-<span class="sd"> Generic class for random dataset splits.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> ds : `pysegcnn.core.dataset.ImageDataset`</span>
-<span class="sd"> An instance of `~pysegcnn.core.dataset.ImageDataset`.</span>
-<span class="sd"> tvratio : `float`</span>
-<span class="sd"> The ratio of training data to validation data, e.g. ``tvratio`` = 0.8</span>
-<span class="sd"> means 80% training, 20% validation.</span>
-<span class="sd"> ttratio : `float`, optional</span>
-<span class="sd"> The ratio of training and validation data to test data, e.g.</span>
-<span class="sd"> ``ttratio`` = 0.6 means 60% for training and validation, 40% for</span>
-<span class="sd"> testing. The default is 1.</span>
-<span class="sd"> seed : `int`, optional</span>
-<span class="sd"> The random seed for reproducibility. The default is 0.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">ds</span><span class="p">,</span> <span class="n">ttratio</span><span class="p">,</span> <span class="n">tvratio</span><span class="p">,</span> <span class="n">seed</span><span class="p">):</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">ds</span><span class="p">)</span>
-
- <span class="c1"># the training, validation and test set ratios</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">ttratio</span> <span class="o">=</span> <span class="n">ttratio</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">tvratio</span> <span class="o">=</span> <span class="n">tvratio</span>
-
- <span class="c1"># the random seed: useful for reproducibility</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">seed</span> <span class="o">=</span> <span class="n">seed</span></div>
-
-
-<div class="viewcode-block" id="RandomTileSplit"><a class="viewcode-back" href="../../source/core.html#core.split.RandomTileSplit">[docs]</a><span class="k">class</span> <span class="nc">RandomTileSplit</span><span class="p">(</span><span class="n">RandomSplit</span><span class="p">):</span>
- <span class="sd">"""Randomly split the dataset.</span>
-
-<span class="sd"> For each scene, the tiles of the scene can be distributed among the</span>
-<span class="sd"> training, validation and test set.</span>
-
-<span class="sd"> Class wrapper for `pysegcnn.core.split.Split.random_tile_split`.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> ds : `pysegcnn.core.dataset.ImageDataset`</span>
-<span class="sd"> An instance of `~pysegcnn.core.dataset.ImageDataset`.</span>
-<span class="sd"> tvratio : `float`</span>
-<span class="sd"> The ratio of training data to validation data, e.g. ``tvratio`` = 0.8</span>
-<span class="sd"> means 80% training, 20% validation.</span>
-<span class="sd"> ttratio : `float`, optional</span>
-<span class="sd"> The ratio of training and validation data to test data, e.g.</span>
-<span class="sd"> ``ttratio`` = 0.6 means 60% for training and validation, 40% for</span>
-<span class="sd"> testing. The default is 1.</span>
-<span class="sd"> seed : `int`, optional</span>
-<span class="sd"> The random seed for reproducibility. The default is 0.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">ds</span><span class="p">,</span> <span class="n">ttratio</span><span class="p">,</span> <span class="n">tvratio</span><span class="p">,</span> <span class="n">seed</span><span class="p">):</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">ds</span><span class="p">,</span> <span class="n">ttratio</span><span class="p">,</span> <span class="n">tvratio</span><span class="p">,</span> <span class="n">seed</span><span class="p">)</span>
-
-<div class="viewcode-block" id="RandomTileSplit.subsets"><a class="viewcode-back" href="../../source/core.html#core.split.RandomTileSplit.subsets">[docs]</a> <span class="k">def</span> <span class="nf">subsets</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Wrap `pysegcnn.core.split.Split.random_tile_split`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> subsets : `dict`</span>
-<span class="sd"> Subset dictionary with keys:</span>
-<span class="sd"> ``'train'``</span>
-<span class="sd"> dictionary containing the training scenes.</span>
-<span class="sd"> ``'valid'``</span>
-<span class="sd"> dictionary containing the validation scenes.</span>
-<span class="sd"> ``'test'``</span>
-<span class="sd"> dictionary containing the test scenes, empty.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="n">random_tile_split</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">ds</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">tvratio</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">ttratio</span><span class="p">,</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">seed</span><span class="p">)</span></div>
-
-<div class="viewcode-block" id="RandomTileSplit.subset_type"><a class="viewcode-back" href="../../source/core.html#core.split.RandomTileSplit.subset_type">[docs]</a> <span class="k">def</span> <span class="nf">subset_type</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Wrap `pysegcnn.core.split.RandomSubset`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> SceneSubset : `pysegcnn.core.split.RandomSubset`</span>
-<span class="sd"> The subset type.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="n">RandomSubset</span></div></div>
-
-
-<div class="viewcode-block" id="RandomSceneSplit"><a class="viewcode-back" href="../../source/core.html#core.split.RandomSceneSplit">[docs]</a><span class="k">class</span> <span class="nc">RandomSceneSplit</span><span class="p">(</span><span class="n">RandomSplit</span><span class="p">):</span>
- <span class="sd">"""Randomly split the dataset.</span>
-
-<span class="sd"> For each scene, all the tiles of the scene are included in either the</span>
-<span class="sd"> training, validation or test set, respectively.</span>
-
-<span class="sd"> Class wrapper for `pysegcnn.core.split.Split.random_scene_split`.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> ds : `pysegcnn.core.dataset.ImageDataset`</span>
-<span class="sd"> An instance of `~pysegcnn.core.dataset.ImageDataset`.</span>
-<span class="sd"> tvratio : `float`</span>
-<span class="sd"> The ratio of training data to validation data, e.g. ``tvratio`` = 0.8</span>
-<span class="sd"> means 80% training, 20% validation.</span>
-<span class="sd"> ttratio : `float`, optional</span>
-<span class="sd"> The ratio of training and validation data to test data, e.g.</span>
-<span class="sd"> ``ttratio`` = 0.6 means 60% for training and validation, 40% for</span>
-<span class="sd"> testing. The default is 1.</span>
-<span class="sd"> seed : `int`, optional</span>
-<span class="sd"> The random seed for reproducibility. The default is 0.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">ds</span><span class="p">,</span> <span class="n">ttratio</span><span class="p">,</span> <span class="n">tvratio</span><span class="p">,</span> <span class="n">seed</span><span class="p">):</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">ds</span><span class="p">,</span> <span class="n">ttratio</span><span class="p">,</span> <span class="n">tvratio</span><span class="p">,</span> <span class="n">seed</span><span class="p">)</span>
-
-<div class="viewcode-block" id="RandomSceneSplit.subsets"><a class="viewcode-back" href="../../source/core.html#core.split.RandomSceneSplit.subsets">[docs]</a> <span class="k">def</span> <span class="nf">subsets</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Wrap `pysegcnn.core.split.Split.random_scene_split`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> subsets : `dict`</span>
-<span class="sd"> Subset dictionary with keys:</span>
-<span class="sd"> ``'train'``</span>
-<span class="sd"> dictionary containing the training scenes.</span>
-<span class="sd"> ``'valid'``</span>
-<span class="sd"> dictionary containing the validation scenes.</span>
-<span class="sd"> ``'test'``</span>
-<span class="sd"> dictionary containing the test scenes, empty.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="n">random_scene_split</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">ds</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">tvratio</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">ttratio</span><span class="p">,</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">seed</span><span class="p">)</span></div>
-
-<div class="viewcode-block" id="RandomSceneSplit.subset_type"><a class="viewcode-back" href="../../source/core.html#core.split.RandomSceneSplit.subset_type">[docs]</a> <span class="k">def</span> <span class="nf">subset_type</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Wrap `pysegcnn.core.split.SceneSubset`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> SceneSubset : `pysegcnn.core.split.SceneSubset`</span>
-<span class="sd"> The subset type.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="n">SceneSubset</span></div></div>
-
-
-<div class="viewcode-block" id="SupportedSplits"><a class="viewcode-back" href="../../source/core.html#core.split.SupportedSplits">[docs]</a><span class="k">class</span> <span class="nc">SupportedSplits</span><span class="p">(</span><span class="n">enum</span><span class="o">.</span><span class="n">Enum</span><span class="p">):</span>
- <span class="sd">"""Names and corresponding classes of the implemented split modes."""</span>
-
- <span class="n">random</span> <span class="o">=</span> <span class="n">RandomTileSplit</span>
- <span class="n">scene</span> <span class="o">=</span> <span class="n">RandomSceneSplit</span>
- <span class="n">date</span> <span class="o">=</span> <span class="n">DateSplit</span></div>
-</pre></div>
-
- </div>
-
- </div>
- <footer>
-
-
- <hr/>
-
- <div role="contentinfo">
- <p>
-
- © Copyright 2020, Daniel Frisinghelli
-
- </p>
- </div>
-
-
-
- Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a
-
- <a href="https://github.com/rtfd/sphinx_rtd_theme">theme</a>
-
- provided by <a href="https://readthedocs.org">Read the Docs</a>.
-
-</footer>
-
- </div>
- </div>
-
- </section>
-
- </div>
-
-
- <script type="text/javascript">
- jQuery(function () {
- SphinxRtdTheme.Navigation.enable(true);
- });
- </script>
-
-
-
-
-
-
-</body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/_modules/core/trainer.html b/docs/_build/html/_modules/core/trainer.html
deleted file mode 100644
index 1822007ba38b704ee90ffb1f66e5a33d6b973067..0000000000000000000000000000000000000000
--- a/docs/_build/html/_modules/core/trainer.html
+++ /dev/null
@@ -1,1760 +0,0 @@
-
-
-<!DOCTYPE html>
-<html class="writer-html5" lang="en" >
-<head>
- <meta charset="utf-8">
-
- <meta name="viewport" content="width=device-width, initial-scale=1.0">
-
- <title>core.trainer — PySegCNN documentation</title>
-
-
-
- <link rel="stylesheet" href="../../_static/css/theme.css" type="text/css" />
- <link rel="stylesheet" href="../../_static/pygments.css" type="text/css" />
-
-
-
-
-
-
-
- <!--[if lt IE 9]>
- <script src="../../_static/js/html5shiv.min.js"></script>
- <![endif]-->
-
-
- <script type="text/javascript" id="documentation_options" data-url_root="../../" src="../../_static/documentation_options.js"></script>
- <script src="../../_static/jquery.js"></script>
- <script src="../../_static/underscore.js"></script>
- <script src="../../_static/doctools.js"></script>
- <script src="../../_static/language_data.js"></script>
-
- <script type="text/javascript" src="../../_static/js/theme.js"></script>
-
-
- <link rel="index" title="Index" href="../../genindex.html" />
- <link rel="search" title="Search" href="../../search.html" />
-</head>
-
-<body class="wy-body-for-nav">
-
-
- <div class="wy-grid-for-nav">
-
- <nav data-toggle="wy-nav-shift" class="wy-nav-side">
- <div class="wy-side-scroll">
- <div class="wy-side-nav-search" >
-
-
-
- <a href="../../index.html" class="icon icon-home" alt="Documentation Home"> PySegCNN
-
-
-
- </a>
-
-
-
-
-
-
-
-<div role="search">
- <form id="rtd-search-form" class="wy-form" action="../../search.html" method="get">
- <input type="text" name="q" placeholder="Search docs" />
- <input type="hidden" name="check_keywords" value="yes" />
- <input type="hidden" name="area" value="default" />
- </form>
-</div>
-
-
- </div>
-
-
- <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
-
-
-
-
-
-
- <ul>
-<li class="toctree-l1"><a class="reference internal" href="../../source/requirements.html">Requirements</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../../source/installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../../source/api.html">API Reference</a></li>
-</ul>
-
-
-
- </div>
-
- </div>
- </nav>
-
- <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
-
-
- <nav class="wy-nav-top" aria-label="top navigation">
-
- <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
- <a href="../../index.html">PySegCNN</a>
-
- </nav>
-
-
- <div class="wy-nav-content">
-
- <div class="rst-content">
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-<div role="navigation" aria-label="breadcrumbs navigation">
-
- <ul class="wy-breadcrumbs">
-
- <li><a href="../../index.html" class="icon icon-home"></a> »</li>
-
- <li><a href="../index.html">Module code</a> »</li>
-
- <li>core.trainer</li>
-
-
- <li class="wy-breadcrumbs-aside">
-
- </li>
-
- </ul>
-
-
- <hr/>
-</div>
- <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
- <div itemprop="articleBody">
-
- <h1>Source code for core.trainer</h1><div class="highlight"><pre>
-<span></span><span class="sd">"""Model configuration and training.</span>
-
-<span class="sd">This module provides an end-to-end framework of dataclasses designed to train</span>
-<span class="sd">segmentation models on image datasets.</span>
-
-<span class="sd">See pysegcnn/main/train.py for a complete walkthrough.</span>
-
-<span class="sd">License</span>
-<span class="sd">-------</span>
-
-<span class="sd"> Copyright (c) 2020 Daniel Frisinghelli</span>
-
-<span class="sd"> This source code is licensed under the GNU General Public License v3.</span>
-
-<span class="sd"> See the LICENSE file in the repository's root directory.</span>
-
-<span class="sd">"""</span>
-
-<span class="c1"># !/usr/bin/env python</span>
-<span class="c1"># -*- coding: utf-8 -*-</span>
-
-<span class="c1"># builtins</span>
-<span class="kn">import</span> <span class="nn">dataclasses</span>
-<span class="kn">import</span> <span class="nn">pathlib</span>
-<span class="kn">import</span> <span class="nn">logging</span>
-<span class="kn">import</span> <span class="nn">datetime</span>
-
-<span class="c1"># externals</span>
-<span class="kn">import</span> <span class="nn">numpy</span> <span class="k">as</span> <span class="nn">np</span>
-<span class="kn">import</span> <span class="nn">torch</span>
-<span class="kn">import</span> <span class="nn">torch.nn</span> <span class="k">as</span> <span class="nn">nn</span>
-<span class="kn">import</span> <span class="nn">torch.nn.functional</span> <span class="k">as</span> <span class="nn">F</span>
-<span class="kn">from</span> <span class="nn">torch.utils.data</span> <span class="kn">import</span> <span class="n">DataLoader</span><span class="p">,</span> <span class="n">Dataset</span>
-<span class="kn">from</span> <span class="nn">torch.optim</span> <span class="kn">import</span> <span class="n">Optimizer</span>
-
-<span class="c1"># locals</span>
-<span class="kn">from</span> <span class="nn">pysegcnn.core.dataset</span> <span class="kn">import</span> <span class="n">SupportedDatasets</span><span class="p">,</span> <span class="n">ImageDataset</span>
-<span class="kn">from</span> <span class="nn">pysegcnn.core.transforms</span> <span class="kn">import</span> <span class="n">Augment</span>
-<span class="kn">from</span> <span class="nn">pysegcnn.core.utils</span> <span class="kn">import</span> <span class="n">img2np</span><span class="p">,</span> <span class="n">item_in_enum</span><span class="p">,</span> <span class="n">accuracy_function</span>
-<span class="kn">from</span> <span class="nn">pysegcnn.core.split</span> <span class="kn">import</span> <span class="n">SupportedSplits</span>
-<span class="kn">from</span> <span class="nn">pysegcnn.core.models</span> <span class="kn">import</span> <span class="p">(</span><span class="n">SupportedModels</span><span class="p">,</span> <span class="n">SupportedOptimizers</span><span class="p">,</span>
- <span class="n">SupportedLossFunctions</span><span class="p">,</span> <span class="n">Network</span><span class="p">)</span>
-<span class="kn">from</span> <span class="nn">pysegcnn.core.layers</span> <span class="kn">import</span> <span class="n">Conv2dSame</span>
-<span class="kn">from</span> <span class="nn">pysegcnn.main.config</span> <span class="kn">import</span> <span class="n">HERE</span>
-
-<span class="c1"># module level logger</span>
-<span class="n">LOGGER</span> <span class="o">=</span> <span class="n">logging</span><span class="o">.</span><span class="n">getLogger</span><span class="p">(</span><span class="vm">__name__</span><span class="p">)</span>
-
-
-<div class="viewcode-block" id="BaseConfig"><a class="viewcode-back" href="../../source/core.html#core.trainer.BaseConfig">[docs]</a><span class="nd">@dataclasses</span><span class="o">.</span><span class="n">dataclass</span>
-<span class="k">class</span> <span class="nc">BaseConfig</span><span class="p">:</span>
- <span class="sd">"""Base `dataclasses.dataclass` for each configuration."""</span>
-
- <span class="k">def</span> <span class="nf">__post_init__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Check the type of each argument.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> TypeError</span>
-<span class="sd"> Raised if the conversion to the specified type of the argument</span>
-<span class="sd"> fails.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># check input types</span>
- <span class="k">for</span> <span class="n">field</span> <span class="ow">in</span> <span class="n">dataclasses</span><span class="o">.</span><span class="n">fields</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="c1"># the value of the current field</span>
- <span class="n">value</span> <span class="o">=</span> <span class="nb">getattr</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">field</span><span class="o">.</span><span class="n">name</span><span class="p">)</span>
-
- <span class="c1"># check whether the value is of the correct type</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">value</span><span class="p">,</span> <span class="n">field</span><span class="o">.</span><span class="n">type</span><span class="p">):</span>
- <span class="c1"># try to convert the value to the correct type</span>
- <span class="k">try</span><span class="p">:</span>
- <span class="nb">setattr</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">field</span><span class="o">.</span><span class="n">name</span><span class="p">,</span> <span class="n">field</span><span class="o">.</span><span class="n">type</span><span class="p">(</span><span class="n">value</span><span class="p">))</span>
- <span class="k">except</span> <span class="ne">TypeError</span><span class="p">:</span>
- <span class="c1"># raise an exception if the conversion fails</span>
- <span class="k">raise</span> <span class="ne">TypeError</span><span class="p">(</span><span class="s1">'Expected </span><span class="si">{}</span><span class="s1"> to be </span><span class="si">{}</span><span class="s1">, got </span><span class="si">{}</span><span class="s1">.'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">field</span><span class="o">.</span><span class="n">name</span><span class="p">,</span> <span class="n">field</span><span class="o">.</span><span class="n">type</span><span class="p">,</span>
- <span class="nb">type</span><span class="p">(</span><span class="n">value</span><span class="p">)))</span></div>
-
-
-<div class="viewcode-block" id="DatasetConfig"><a class="viewcode-back" href="../../source/core.html#core.trainer.DatasetConfig">[docs]</a><span class="nd">@dataclasses</span><span class="o">.</span><span class="n">dataclass</span>
-<span class="k">class</span> <span class="nc">DatasetConfig</span><span class="p">(</span><span class="n">BaseConfig</span><span class="p">):</span>
- <span class="sd">"""Dataset configuration class.</span>
-
-<span class="sd"> Instanciate a dataset.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> dataset_name : `str`</span>
-<span class="sd"> The name of the dataset.</span>
-<span class="sd"> root_dir : `pathlib.Path`</span>
-<span class="sd"> The root directory, path to the dataset.</span>
-<span class="sd"> bands : `list` [`str`]</span>
-<span class="sd"> A list of the spectral bands to use.</span>
-<span class="sd"> tile_size : `int`</span>
-<span class="sd"> The size of the tiles. Each scene is divided into square tiles of shape</span>
-<span class="sd"> (tile_size, tile_size).</span>
-<span class="sd"> gt_pattern : `str`</span>
-<span class="sd"> A pattern to match the ground truth naming convention. All directories</span>
-<span class="sd"> and subdirectories in ``root_dir`` are searched for files matching</span>
-<span class="sd"> ``gt_pattern``.</span>
-<span class="sd"> seed : `int`</span>
-<span class="sd"> The random seed. Used to split the dataset into training, validation</span>
-<span class="sd"> and test set. Useful for reproducibility. The default is 0.</span>
-<span class="sd"> sort : `bool`, optional</span>
-<span class="sd"> Whether to chronologically sort the samples. Useful for time series</span>
-<span class="sd"> data. The default is False.</span>
-<span class="sd"> transforms : `list` [`pysegcnn.core.split.Augment`], optional</span>
-<span class="sd"> List of `pysegcnn.core.split.Augment` instances. Each item in</span>
-<span class="sd"> ``transforms`` generates a distinct transformed version of the dataset.</span>
-<span class="sd"> The total dataset is composed of the original untransformed dataset</span>
-<span class="sd"> together with each transformed version of it.</span>
-<span class="sd"> If ``transforms`` = [], only the original dataset is used.</span>
-<span class="sd"> The default is [].</span>
-<span class="sd"> pad : `bool`, optional</span>
-<span class="sd"> Whether to center pad the input image. Set ``pad`` = True, if the</span>
-<span class="sd"> images are not evenly divisible by the ``tile_size``. The image data is</span>
-<span class="sd"> padded with a constant padding value of zero. For each image, the</span>
-<span class="sd"> corresponding ground truth image is padded with a "no data" label.</span>
-<span class="sd"> The default is False.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="n">dataset_name</span><span class="p">:</span> <span class="nb">str</span>
- <span class="n">root_dir</span><span class="p">:</span> <span class="n">pathlib</span><span class="o">.</span><span class="n">Path</span>
- <span class="n">bands</span><span class="p">:</span> <span class="nb">list</span>
- <span class="n">tile_size</span><span class="p">:</span> <span class="nb">int</span>
- <span class="n">gt_pattern</span><span class="p">:</span> <span class="nb">str</span>
- <span class="n">seed</span><span class="p">:</span> <span class="nb">int</span>
- <span class="n">sort</span><span class="p">:</span> <span class="nb">bool</span> <span class="o">=</span> <span class="kc">False</span>
- <span class="n">transforms</span><span class="p">:</span> <span class="nb">list</span> <span class="o">=</span> <span class="n">dataclasses</span><span class="o">.</span><span class="n">field</span><span class="p">(</span><span class="n">default_factory</span><span class="o">=</span><span class="nb">list</span><span class="p">)</span>
- <span class="n">pad</span><span class="p">:</span> <span class="nb">bool</span> <span class="o">=</span> <span class="kc">False</span>
-
- <span class="k">def</span> <span class="nf">__post_init__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Check the type of each argument.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> ValueError</span>
-<span class="sd"> Raised if ``dataset_name`` is not supported.</span>
-<span class="sd"> FileNotFoundError</span>
-<span class="sd"> Raised if ``root_dir`` does not exist.</span>
-<span class="sd"> TypeError</span>
-<span class="sd"> Raised if not each item in ``transforms`` is an instance of</span>
-<span class="sd"> `pysegcnn.core.split.Augment` in case ``transforms`` is not empty.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># check input types</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="n">__post_init__</span><span class="p">()</span>
-
- <span class="c1"># check whether the dataset is currently supported</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">dataset_class</span> <span class="o">=</span> <span class="n">item_in_enum</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">dataset_name</span><span class="p">,</span> <span class="n">SupportedDatasets</span><span class="p">)</span>
-
- <span class="c1"># check whether the root directory exists</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="bp">self</span><span class="o">.</span><span class="n">root_dir</span><span class="o">.</span><span class="n">exists</span><span class="p">():</span>
- <span class="k">raise</span> <span class="ne">FileNotFoundError</span><span class="p">(</span><span class="s1">'</span><span class="si">{}</span><span class="s1"> does not exist.'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">root_dir</span><span class="p">))</span>
-
- <span class="c1"># check whether the transformations inherit from the correct class</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="nb">all</span><span class="p">([</span><span class="nb">isinstance</span><span class="p">(</span><span class="n">t</span><span class="p">,</span> <span class="n">Augment</span><span class="p">)</span> <span class="k">for</span> <span class="n">t</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">transforms</span> <span class="k">if</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">transforms</span><span class="p">]):</span>
- <span class="k">raise</span> <span class="ne">TypeError</span><span class="p">(</span><span class="s1">'Each transformation is expected to be an instance'</span>
- <span class="s1">' of </span><span class="si">{}</span><span class="s1">.'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="s1">'.'</span><span class="o">.</span><span class="n">join</span><span class="p">([</span><span class="n">Augment</span><span class="o">.</span><span class="vm">__module__</span><span class="p">,</span>
- <span class="n">Augment</span><span class="o">.</span><span class="vm">__name__</span><span class="p">])))</span>
-
-<div class="viewcode-block" id="DatasetConfig.init_dataset"><a class="viewcode-back" href="../../source/core.html#core.trainer.DatasetConfig.init_dataset">[docs]</a> <span class="k">def</span> <span class="nf">init_dataset</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Instanciate the dataset.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> dataset : `pysegcnn.core.dataset.ImageDataset`</span>
-<span class="sd"> An instance of `pysegcnn.core.dataset.ImageDataset`.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># instanciate the dataset</span>
- <span class="n">dataset</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">dataset_class</span><span class="p">(</span>
- <span class="n">root_dir</span><span class="o">=</span><span class="nb">str</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">root_dir</span><span class="p">),</span>
- <span class="n">use_bands</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">bands</span><span class="p">,</span>
- <span class="n">tile_size</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">tile_size</span><span class="p">,</span>
- <span class="n">seed</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">seed</span><span class="p">,</span>
- <span class="n">sort</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">sort</span><span class="p">,</span>
- <span class="n">transforms</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">transforms</span><span class="p">,</span>
- <span class="n">pad</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">pad</span><span class="p">,</span>
- <span class="n">gt_pattern</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">gt_pattern</span>
- <span class="p">)</span>
-
- <span class="k">return</span> <span class="n">dataset</span></div></div>
-
-
-<div class="viewcode-block" id="SplitConfig"><a class="viewcode-back" href="../../source/core.html#core.trainer.SplitConfig">[docs]</a><span class="nd">@dataclasses</span><span class="o">.</span><span class="n">dataclass</span>
-<span class="k">class</span> <span class="nc">SplitConfig</span><span class="p">(</span><span class="n">BaseConfig</span><span class="p">):</span>
- <span class="sd">"""Dataset split configuration class.</span>
-
-<span class="sd"> Split a dataset into training, validation and test set.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> split_mode : `str`</span>
-<span class="sd"> The mode to split the dataset.</span>
-<span class="sd"> ttratio : `float`</span>
-<span class="sd"> The ratio of training and validation data to test data, e.g.</span>
-<span class="sd"> ``ttratio`` = 0.6 means 60% for training and validation, 40% for</span>
-<span class="sd"> testing.</span>
-<span class="sd"> tvratio : `float`</span>
-<span class="sd"> The ratio of training data to validation data, e.g. ``tvratio`` = 0.8</span>
-<span class="sd"> means 80% training, 20% validation.</span>
-<span class="sd"> date : `str`, optional</span>
-<span class="sd"> A date. Used if ``split_mode`` = 'date'. The default is 'yyyymmdd'.</span>
-<span class="sd"> dateformat : `str`, optional</span>
-<span class="sd"> The format of ``date``. ``dateformat`` is used by</span>
-<span class="sd"> `datetime.datetime.strptime' to parse ``date`` to a `datetime.datetime`</span>
-<span class="sd"> object. The default is '%Y%m%d'.</span>
-<span class="sd"> drop : `float`, optional</span>
-<span class="sd"> Whether to drop samples (during training only) with a fraction of</span>
-<span class="sd"> pixels equal to the constant padding value >= ``drop``. ``drop`` = 0</span>
-<span class="sd"> means, do not drop any samples. The default is 0.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="n">split_mode</span><span class="p">:</span> <span class="nb">str</span>
- <span class="n">ttratio</span><span class="p">:</span> <span class="nb">float</span>
- <span class="n">tvratio</span><span class="p">:</span> <span class="nb">float</span>
- <span class="n">date</span><span class="p">:</span> <span class="nb">str</span> <span class="o">=</span> <span class="s1">'yyyymmdd'</span>
- <span class="n">dateformat</span><span class="p">:</span> <span class="nb">str</span> <span class="o">=</span> <span class="s1">'%Y%m</span><span class="si">%d</span><span class="s1">'</span>
- <span class="n">drop</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mi">0</span>
-
- <span class="k">def</span> <span class="nf">__post_init__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Check the type of each argument.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> ValueError</span>
-<span class="sd"> Raised if ``split_mode`` is not supported.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># check input types</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="n">__post_init__</span><span class="p">()</span>
-
- <span class="c1"># check if the split mode is valid</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">split_class</span> <span class="o">=</span> <span class="n">item_in_enum</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">split_mode</span><span class="p">,</span> <span class="n">SupportedSplits</span><span class="p">)</span>
-
- <span class="nd">@staticmethod</span>
- <span class="k">def</span> <span class="nf">_drop_samples</span><span class="p">(</span><span class="n">ds</span><span class="p">,</span> <span class="n">drop_threshold</span><span class="o">=</span><span class="mi">1</span><span class="p">):</span>
- <span class="sd">"""Drop samples with a fraction of pixels equal to the padding value.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> ds : `pysegcnn.core.split.RandomSubset` or</span>
-<span class="sd"> `pysegcnn.core.split.SceneSubset`.</span>
-<span class="sd"> An instance of `pysegcnn.core.split.RandomSubset` or</span>
-<span class="sd"> `pysegcnn.core.split.SceneSubset`.</span>
-<span class="sd"> drop_threshold : `float`, optional</span>
-<span class="sd"> The threshold above which samples are dropped. ``drop_threshold`` =</span>
-<span class="sd"> 1 means a sample is dropped, if all pixels are equal to the padding</span>
-<span class="sd"> value. ``drop_threshold`` = 0.8 means, drop a sample if 80% of the</span>
-<span class="sd"> pixels are equal to the padding value, etc. The default is 1.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> dropped : `list` [`dict`]</span>
-<span class="sd"> List of the dropped samples.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># iterate over the scenes returned by self.compose_scenes()</span>
- <span class="n">dropped</span> <span class="o">=</span> <span class="p">[]</span>
- <span class="k">for</span> <span class="n">pos</span><span class="p">,</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">(</span><span class="n">ds</span><span class="o">.</span><span class="n">indices</span><span class="p">):</span>
-
- <span class="c1"># the current scene</span>
- <span class="n">s</span> <span class="o">=</span> <span class="n">ds</span><span class="o">.</span><span class="n">dataset</span><span class="o">.</span><span class="n">scenes</span><span class="p">[</span><span class="n">i</span><span class="p">]</span>
-
- <span class="c1"># the current tile in the ground truth</span>
- <span class="n">tile_gt</span> <span class="o">=</span> <span class="n">img2np</span><span class="p">(</span><span class="n">s</span><span class="p">[</span><span class="s1">'gt'</span><span class="p">],</span> <span class="n">ds</span><span class="o">.</span><span class="n">dataset</span><span class="o">.</span><span class="n">tile_size</span><span class="p">,</span> <span class="n">s</span><span class="p">[</span><span class="s1">'tile'</span><span class="p">],</span>
- <span class="n">ds</span><span class="o">.</span><span class="n">dataset</span><span class="o">.</span><span class="n">pad</span><span class="p">,</span> <span class="n">ds</span><span class="o">.</span><span class="n">dataset</span><span class="o">.</span><span class="n">cval</span><span class="p">)</span>
-
- <span class="c1"># percent of pixels equal to the constant padding value</span>
- <span class="n">npixels</span> <span class="o">=</span> <span class="p">(</span><span class="n">tile_gt</span><span class="p">[</span><span class="n">tile_gt</span> <span class="o">==</span> <span class="n">ds</span><span class="o">.</span><span class="n">dataset</span><span class="o">.</span><span class="n">cval</span><span class="p">]</span><span class="o">.</span><span class="n">size</span> <span class="o">/</span> <span class="n">tile_gt</span><span class="o">.</span><span class="n">size</span><span class="p">)</span>
-
- <span class="c1"># drop samples where npixels >= self.drop</span>
- <span class="k">if</span> <span class="n">npixels</span> <span class="o">>=</span> <span class="n">drop_threshold</span><span class="p">:</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Skipping scene </span><span class="si">{}</span><span class="s1">, tile </span><span class="si">{}</span><span class="s1">: </span><span class="si">{:.2f}</span><span class="s1">% padded pixels'</span>
- <span class="s1">' ...'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">s</span><span class="p">[</span><span class="s1">'id'</span><span class="p">],</span> <span class="n">s</span><span class="p">[</span><span class="s1">'tile'</span><span class="p">],</span> <span class="n">npixels</span> <span class="o">*</span> <span class="mi">100</span><span class="p">))</span>
- <span class="n">dropped</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">s</span><span class="p">)</span>
- <span class="n">_</span> <span class="o">=</span> <span class="n">ds</span><span class="o">.</span><span class="n">indices</span><span class="o">.</span><span class="n">pop</span><span class="p">(</span><span class="n">pos</span><span class="p">)</span>
-
- <span class="k">return</span> <span class="n">dropped</span>
-
-<div class="viewcode-block" id="SplitConfig.train_val_test_split"><a class="viewcode-back" href="../../source/core.html#core.trainer.SplitConfig.train_val_test_split">[docs]</a> <span class="k">def</span> <span class="nf">train_val_test_split</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">ds</span><span class="p">):</span>
- <span class="sd">"""Split ``ds`` into training, validation and test set.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> ds : `pysegcnn.core.dataset.ImageDataset`</span>
-<span class="sd"> An instance of `pysegcnn.core.dataset.ImageDataset`.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> TypeError</span>
-<span class="sd"> Raised if ``ds`` is not an instance of</span>
-<span class="sd"> `pysegcnn.core.dataset.ImageDataset`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> train_ds : `pysegcnn.core.split.RandomSubset` or</span>
-<span class="sd"> `pysegcnn.core.split.SceneSubset`.</span>
-<span class="sd"> The training set.</span>
-<span class="sd"> valid_ds : `pysegcnn.core.split.RandomSubset` or</span>
-<span class="sd"> `pysegcnn.core.split.SceneSubset`.</span>
-<span class="sd"> The validation set.</span>
-<span class="sd"> test_ds : `pysegcnn.core.split.RandomSubset` or</span>
-<span class="sd"> `pysegcnn.core.split.SceneSubset`.</span>
-<span class="sd"> The test set.</span>
-
-<span class="sd"> """</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">ds</span><span class="p">,</span> <span class="n">ImageDataset</span><span class="p">):</span>
- <span class="k">raise</span> <span class="ne">TypeError</span><span class="p">(</span><span class="s1">'Expected "ds" to be </span><span class="si">{}</span><span class="s1">.'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="s1">'.'</span><span class="o">.</span><span class="n">join</span><span class="p">([</span><span class="n">ImageDataset</span><span class="o">.</span><span class="vm">__module__</span><span class="p">,</span>
- <span class="n">ImageDataset</span><span class="o">.</span><span class="vm">__name__</span><span class="p">])))</span>
-
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">split_mode</span> <span class="o">==</span> <span class="s1">'random'</span> <span class="ow">or</span> <span class="bp">self</span><span class="o">.</span><span class="n">split_mode</span> <span class="o">==</span> <span class="s1">'scene'</span><span class="p">:</span>
- <span class="n">subset</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">split_class</span><span class="p">(</span><span class="n">ds</span><span class="p">,</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">ttratio</span><span class="p">,</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">tvratio</span><span class="p">,</span>
- <span class="n">ds</span><span class="o">.</span><span class="n">seed</span><span class="p">)</span>
-
- <span class="k">else</span><span class="p">:</span>
- <span class="n">subset</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">split_class</span><span class="p">(</span><span class="n">ds</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">date</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">dateformat</span><span class="p">)</span>
-
- <span class="c1"># the training, validation and test dataset</span>
- <span class="n">train_ds</span><span class="p">,</span> <span class="n">valid_ds</span><span class="p">,</span> <span class="n">test_ds</span> <span class="o">=</span> <span class="n">subset</span><span class="o">.</span><span class="n">split</span><span class="p">()</span>
-
- <span class="c1"># whether to drop training samples with a fraction of pixels equal to</span>
- <span class="c1"># the constant padding value cval >= drop</span>
- <span class="k">if</span> <span class="n">ds</span><span class="o">.</span><span class="n">pad</span> <span class="ow">and</span> <span class="bp">self</span><span class="o">.</span><span class="n">drop</span> <span class="o">></span> <span class="mi">0</span><span class="p">:</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">dropped</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">_drop_samples</span><span class="p">(</span><span class="n">train_ds</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">drop</span><span class="p">)</span>
-
- <span class="k">return</span> <span class="n">train_ds</span><span class="p">,</span> <span class="n">valid_ds</span><span class="p">,</span> <span class="n">test_ds</span></div>
-
-<div class="viewcode-block" id="SplitConfig.dataloaders"><a class="viewcode-back" href="../../source/core.html#core.trainer.SplitConfig.dataloaders">[docs]</a> <span class="nd">@staticmethod</span>
- <span class="k">def</span> <span class="nf">dataloaders</span><span class="p">(</span><span class="o">*</span><span class="n">args</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
- <span class="sd">"""Build `torch.utils.data.DataLoader` instances.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> *args : `list` [`torch.utils.data.Dataset`]</span>
-<span class="sd"> List of instances of `torch.utils.data.Dataset`.</span>
-<span class="sd"> **kwargs</span>
-<span class="sd"> Additional keyword arguments passed to</span>
-<span class="sd"> `torch.utils.data.DataLoader`.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> TypeError</span>
-<span class="sd"> Raised if not each item in ``args`` is an instance of</span>
-<span class="sd"> `torch.utils.data.Dataset`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> loaders : `list` [`torch.utils.data.DataLoader`]</span>
-<span class="sd"> List of instances of `torch.utils.data.DataLoader`. If an instance</span>
-<span class="sd"> of `torch.utils.data.Dataset` in ``args`` is empty, `None` is</span>
-<span class="sd"> appended to ``loaders`` instead of an instance of</span>
-<span class="sd"> `torch.utils.data.DataLoader`.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># check whether each dataset in args has the correct type</span>
- <span class="n">loaders</span> <span class="o">=</span> <span class="p">[]</span>
- <span class="k">for</span> <span class="n">ds</span> <span class="ow">in</span> <span class="n">args</span><span class="p">:</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">ds</span><span class="p">,</span> <span class="n">Dataset</span><span class="p">):</span>
- <span class="k">raise</span> <span class="ne">TypeError</span><span class="p">(</span><span class="s1">'Expected </span><span class="si">{}</span><span class="s1">, got </span><span class="si">{}</span><span class="s1">.'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="nb">repr</span><span class="p">(</span><span class="n">Dataset</span><span class="p">),</span> <span class="nb">type</span><span class="p">(</span><span class="n">ds</span><span class="p">)))</span>
-
- <span class="c1"># check if the dataset is not empty</span>
- <span class="k">if</span> <span class="nb">len</span><span class="p">(</span><span class="n">ds</span><span class="p">)</span> <span class="o">></span> <span class="mi">0</span><span class="p">:</span>
- <span class="c1"># build the dataloader</span>
- <span class="n">loader</span> <span class="o">=</span> <span class="n">DataLoader</span><span class="p">(</span><span class="n">ds</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="n">loader</span> <span class="o">=</span> <span class="kc">None</span>
- <span class="n">loaders</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">loader</span><span class="p">)</span>
-
- <span class="k">return</span> <span class="n">loaders</span></div></div>
-
-
-<div class="viewcode-block" id="ModelConfig"><a class="viewcode-back" href="../../source/core.html#core.trainer.ModelConfig">[docs]</a><span class="nd">@dataclasses</span><span class="o">.</span><span class="n">dataclass</span>
-<span class="k">class</span> <span class="nc">ModelConfig</span><span class="p">(</span><span class="n">BaseConfig</span><span class="p">):</span>
- <span class="sd">"""Model configuration class.</span>
-
-<span class="sd"> Instanciate a (pretrained) model.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> model_name : `str`</span>
-<span class="sd"> The name of the model.</span>
-<span class="sd"> filters : `list` [`int`]</span>
-<span class="sd"> List of input channels to the convolutional layers.</span>
-<span class="sd"> torch_seed : `int`</span>
-<span class="sd"> The random seed to initialize the model weights.</span>
-<span class="sd"> Useful for reproducibility.</span>
-<span class="sd"> optim_name : `str`</span>
-<span class="sd"> The name of the optimizer to update the model weights.</span>
-<span class="sd"> loss_name : `str`</span>
-<span class="sd"> The name of the loss function measuring the model error.</span>
-<span class="sd"> skip_connection : `bool`, optional</span>
-<span class="sd"> Whether to apply skip connections. The defaul is True.</span>
-<span class="sd"> kwargs: `dict`, optional</span>
-<span class="sd"> The configuration for each convolution in the model. The default is</span>
-<span class="sd"> {'kernel_size': 3, 'stride': 1, 'dilation': 1}.</span>
-<span class="sd"> batch_size : `int`, optional</span>
-<span class="sd"> The model batch size. Determines the number of samples to process</span>
-<span class="sd"> before updating the model weights. The default is 64.</span>
-<span class="sd"> checkpoint : `bool`, optional</span>
-<span class="sd"> Whether to resume training from an existing model checkpoint. The</span>
-<span class="sd"> default is False.</span>
-<span class="sd"> transfer : `bool`, optional</span>
-<span class="sd"> Whether to use a model for transfer learning on a new dataset. If True,</span>
-<span class="sd"> the model architecture of ``pretrained_model`` is adjusted to a new</span>
-<span class="sd"> dataset. The default is False.</span>
-<span class="sd"> pretrained_model : `str`, optional</span>
-<span class="sd"> The name of the pretrained model to use for transfer learning.</span>
-<span class="sd"> The default is ''.</span>
-<span class="sd"> lr : `float`, optional</span>
-<span class="sd"> The learning rate used by the gradient descent algorithm.</span>
-<span class="sd"> The default is 0.001.</span>
-<span class="sd"> early_stop : `bool`, optional</span>
-<span class="sd"> Whether to apply `early stopping`_. The default is False.</span>
-<span class="sd"> mode : `str`, optional</span>
-<span class="sd"> The mode of the early stopping. Depends on the metric measuring</span>
-<span class="sd"> performance. When using model loss as metric, use ``mode`` = 'min',</span>
-<span class="sd"> however, when using accuracy as metric, use ``mode`` = 'max'. For now,</span>
-<span class="sd"> only ``mode`` = 'max' is supported. Only used if ``early_stop`` = True.</span>
-<span class="sd"> The default is 'max'.</span>
-<span class="sd"> delta : `float`, optional</span>
-<span class="sd"> Minimum change in early stopping metric to be considered as an</span>
-<span class="sd"> improvement. Only used if ``early_stop`` = True. The default is 0.</span>
-<span class="sd"> patience : `int`, optional</span>
-<span class="sd"> The number of epochs to wait for an improvement in the early stopping</span>
-<span class="sd"> metric. If the model does not improve over more than ``patience``</span>
-<span class="sd"> epochs, quit training. Only used if ``early_stop`` = True.</span>
-<span class="sd"> The default is 10.</span>
-<span class="sd"> epochs : `int`, optional</span>
-<span class="sd"> The maximum number of epochs to train. The default is 50.</span>
-<span class="sd"> nthreads : `int`, optional</span>
-<span class="sd"> The number of cpu threads to use during training. The default is</span>
-<span class="sd"> torch.get_num_threads().</span>
-<span class="sd"> save : `bool`, optional</span>
-<span class="sd"> Whether to save the model state to disk. Model states are saved in</span>
-<span class="sd"> pysegcnn/main/_models. The default is True.</span>
-
-<span class="sd"> .. _early stopping:</span>
-<span class="sd"> https://en.wikipedia.org/wiki/Early_stopping</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="n">model_name</span><span class="p">:</span> <span class="nb">str</span>
- <span class="n">filters</span><span class="p">:</span> <span class="nb">list</span>
- <span class="n">torch_seed</span><span class="p">:</span> <span class="nb">int</span>
- <span class="n">optim_name</span><span class="p">:</span> <span class="nb">str</span>
- <span class="n">loss_name</span><span class="p">:</span> <span class="nb">str</span>
- <span class="n">skip_connection</span><span class="p">:</span> <span class="nb">bool</span> <span class="o">=</span> <span class="kc">True</span>
- <span class="n">kwargs</span><span class="p">:</span> <span class="nb">dict</span> <span class="o">=</span> <span class="n">dataclasses</span><span class="o">.</span><span class="n">field</span><span class="p">(</span>
- <span class="n">default_factory</span><span class="o">=</span><span class="k">lambda</span><span class="p">:</span> <span class="p">{</span><span class="s1">'kernel_size'</span><span class="p">:</span> <span class="mi">3</span><span class="p">,</span> <span class="s1">'stride'</span><span class="p">:</span> <span class="mi">1</span><span class="p">,</span> <span class="s1">'dilation'</span><span class="p">:</span> <span class="mi">1</span><span class="p">})</span>
- <span class="n">batch_size</span><span class="p">:</span> <span class="nb">int</span> <span class="o">=</span> <span class="mi">64</span>
- <span class="n">checkpoint</span><span class="p">:</span> <span class="nb">bool</span> <span class="o">=</span> <span class="kc">False</span>
- <span class="n">transfer</span><span class="p">:</span> <span class="nb">bool</span> <span class="o">=</span> <span class="kc">False</span>
- <span class="n">pretrained_model</span><span class="p">:</span> <span class="nb">str</span> <span class="o">=</span> <span class="s1">''</span>
- <span class="n">lr</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mf">0.001</span>
- <span class="n">early_stop</span><span class="p">:</span> <span class="nb">bool</span> <span class="o">=</span> <span class="kc">False</span>
- <span class="n">mode</span><span class="p">:</span> <span class="nb">str</span> <span class="o">=</span> <span class="s1">'max'</span>
- <span class="n">delta</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mi">0</span>
- <span class="n">patience</span><span class="p">:</span> <span class="nb">int</span> <span class="o">=</span> <span class="mi">10</span>
- <span class="n">epochs</span><span class="p">:</span> <span class="nb">int</span> <span class="o">=</span> <span class="mi">50</span>
- <span class="n">nthreads</span><span class="p">:</span> <span class="nb">int</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">get_num_threads</span><span class="p">()</span>
- <span class="n">save</span><span class="p">:</span> <span class="nb">bool</span> <span class="o">=</span> <span class="kc">True</span>
-
- <span class="k">def</span> <span class="nf">__post_init__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Check the type of each argument.</span>
-
-<span class="sd"> Configure path to save model state.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> ValueError</span>
-<span class="sd"> Raised if the model ``model_name``, the optimizer ``optim_name`` or</span>
-<span class="sd"> the loss function ``loss_name`` is not supported.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># check input types</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="n">__post_init__</span><span class="p">()</span>
-
- <span class="c1"># check whether the model is currently supported</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">model_class</span> <span class="o">=</span> <span class="n">item_in_enum</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">model_name</span><span class="p">,</span> <span class="n">SupportedModels</span><span class="p">)</span>
-
- <span class="c1"># check whether the optimizer is currently supported</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">optim_class</span> <span class="o">=</span> <span class="n">item_in_enum</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">optim_name</span><span class="p">,</span> <span class="n">SupportedOptimizers</span><span class="p">)</span>
-
- <span class="c1"># check whether the loss function is currently supported</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">loss_class</span> <span class="o">=</span> <span class="n">item_in_enum</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">loss_name</span><span class="p">,</span> <span class="n">SupportedLossFunctions</span><span class="p">)</span>
-
- <span class="c1"># path to model states</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">state_path</span> <span class="o">=</span> <span class="n">pathlib</span><span class="o">.</span><span class="n">Path</span><span class="p">(</span><span class="n">HERE</span><span class="p">)</span><span class="o">.</span><span class="n">joinpath</span><span class="p">(</span><span class="s1">'_models/'</span><span class="p">)</span>
-
- <span class="c1"># path to pretrained model</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">pretrained_path</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">state_path</span><span class="o">.</span><span class="n">joinpath</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">pretrained_model</span><span class="p">)</span>
-
-<div class="viewcode-block" id="ModelConfig.init_optimizer"><a class="viewcode-back" href="../../source/core.html#core.trainer.ModelConfig.init_optimizer">[docs]</a> <span class="k">def</span> <span class="nf">init_optimizer</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">model</span><span class="p">):</span>
- <span class="sd">"""Instanciate the optimizer.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> model : `torch.nn.Module`</span>
-<span class="sd"> An instance of `torch.nn.Module`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> optimizer : `torch.optim.Optimizer`</span>
-<span class="sd"> An instance of `torch.optim.Optimizer`.</span>
-
-<span class="sd"> """</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Optimizer: </span><span class="si">{}</span><span class="s1">.'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="nb">repr</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">optim_class</span><span class="p">)))</span>
-
- <span class="c1"># initialize the optimizer for the specified model</span>
- <span class="n">optimizer</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">optim_class</span><span class="p">(</span><span class="n">model</span><span class="o">.</span><span class="n">parameters</span><span class="p">(),</span> <span class="bp">self</span><span class="o">.</span><span class="n">lr</span><span class="p">)</span>
-
- <span class="k">return</span> <span class="n">optimizer</span></div>
-
-<div class="viewcode-block" id="ModelConfig.init_loss_function"><a class="viewcode-back" href="../../source/core.html#core.trainer.ModelConfig.init_loss_function">[docs]</a> <span class="k">def</span> <span class="nf">init_loss_function</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Instanciate the loss function.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> loss_function : `torch.nn.Module`</span>
-<span class="sd"> An instance of `torch.nn.Module`.</span>
-
-<span class="sd"> """</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Loss function: </span><span class="si">{}</span><span class="s1">.'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="nb">repr</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">loss_class</span><span class="p">)))</span>
-
- <span class="c1"># instanciate the loss function</span>
- <span class="n">loss_function</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">loss_class</span><span class="p">()</span>
-
- <span class="k">return</span> <span class="n">loss_function</span></div>
-
-<div class="viewcode-block" id="ModelConfig.init_model"><a class="viewcode-back" href="../../source/core.html#core.trainer.ModelConfig.init_model">[docs]</a> <span class="k">def</span> <span class="nf">init_model</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">ds</span><span class="p">,</span> <span class="n">state_file</span><span class="p">):</span>
- <span class="sd">"""Instanciate the model and the optimizer.</span>
-
-<span class="sd"> If the model checkpoint ``state_file`` exists, the pretrained model and</span>
-<span class="sd"> optimizer states are loaded, otherwise the model and the optimizer are</span>
-<span class="sd"> initialized from scratch.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> ds : `pysegcnn.core.dataset.ImageDataset`</span>
-<span class="sd"> An instance of `pysegcnn.core.dataset.ImageDataset`.</span>
-<span class="sd"> state_file : `pathlib.Path`</span>
-<span class="sd"> Path to a model checkpoint.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> model : `pysegcnn.core.models.Network`</span>
-<span class="sd"> An instance of `pysegcnn.core.models.Network`.</span>
-<span class="sd"> optimizer : `torch.optim.Optimizer`</span>
-<span class="sd"> An instance of `torch.optim.Optimizer`.</span>
-<span class="sd"> checkpoint_state : `dict` [`str`, `numpy.ndarray`]</span>
-<span class="sd"> If the model checkpoint ``state_file`` exists, ``checkpoint_state``</span>
-<span class="sd"> has keys:</span>
-<span class="sd"> ``'ta'``</span>
-<span class="sd"> The accuracy on the training set (`numpy.ndarray`).</span>
-<span class="sd"> ``'tl'``</span>
-<span class="sd"> The loss on the training set (`numpy.ndarray`).</span>
-<span class="sd"> ``'va'``</span>
-<span class="sd"> The accuracy on the validation set (`numpy.ndarray`).</span>
-<span class="sd"> ``'vl'``</span>
-<span class="sd"> The loss on the validation set (`numpy.ndarray`).</span>
-
-<span class="sd"> """</span>
- <span class="c1"># write an initialization string to the log file</span>
- <span class="n">LogConfig</span><span class="o">.</span><span class="n">init_log</span><span class="p">(</span><span class="s1">'</span><span class="si">{}</span><span class="s1">: Initializing model run. '</span><span class="p">)</span>
-
- <span class="c1"># case (1): build a new model</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="bp">self</span><span class="o">.</span><span class="n">transfer</span><span class="p">:</span>
-
- <span class="c1"># set the random seed for reproducibility</span>
- <span class="n">torch</span><span class="o">.</span><span class="n">manual_seed</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">torch_seed</span><span class="p">)</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Initializing model: </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">state_file</span><span class="o">.</span><span class="n">name</span><span class="p">))</span>
-
- <span class="c1"># instanciate the model</span>
- <span class="n">model</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">model_class</span><span class="p">(</span>
- <span class="n">in_channels</span><span class="o">=</span><span class="nb">len</span><span class="p">(</span><span class="n">ds</span><span class="o">.</span><span class="n">use_bands</span><span class="p">),</span>
- <span class="n">nclasses</span><span class="o">=</span><span class="nb">len</span><span class="p">(</span><span class="n">ds</span><span class="o">.</span><span class="n">labels</span><span class="p">),</span>
- <span class="n">filters</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">filters</span><span class="p">,</span>
- <span class="n">skip</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">skip_connection</span><span class="p">,</span>
- <span class="o">**</span><span class="bp">self</span><span class="o">.</span><span class="n">kwargs</span><span class="p">)</span>
-
- <span class="c1"># case (2): load a pretrained model for transfer learning</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="c1"># load pretrained model</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Loading pretrained model for transfer learning from: '</span>
- <span class="s1">'</span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">pretrained_path</span><span class="p">))</span>
- <span class="n">model</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">transfer_model</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">pretrained_path</span><span class="p">,</span> <span class="n">ds</span><span class="p">)</span>
-
- <span class="c1"># initialize the optimizer</span>
- <span class="n">optimizer</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">init_optimizer</span><span class="p">(</span><span class="n">model</span><span class="p">)</span>
-
- <span class="c1"># whether to resume training from an existing model checkpoint</span>
- <span class="n">checkpoint_state</span> <span class="o">=</span> <span class="p">{}</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">checkpoint</span><span class="p">:</span>
- <span class="n">model</span><span class="p">,</span> <span class="n">optimizer</span><span class="p">,</span> <span class="n">checkpoint_state</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">load_checkpoint</span><span class="p">(</span>
- <span class="n">model</span><span class="p">,</span> <span class="n">optimizer</span><span class="p">,</span> <span class="n">state_file</span><span class="p">)</span>
-
- <span class="k">return</span> <span class="n">model</span><span class="p">,</span> <span class="n">optimizer</span><span class="p">,</span> <span class="n">checkpoint_state</span></div>
-
-<div class="viewcode-block" id="ModelConfig.load_checkpoint"><a class="viewcode-back" href="../../source/core.html#core.trainer.ModelConfig.load_checkpoint">[docs]</a> <span class="nd">@staticmethod</span>
- <span class="k">def</span> <span class="nf">load_checkpoint</span><span class="p">(</span><span class="n">model</span><span class="p">,</span> <span class="n">optimizer</span><span class="p">,</span> <span class="n">state_file</span><span class="p">):</span>
- <span class="sd">"""Load an existing model checkpoint.</span>
-
-<span class="sd"> If the model checkpoint ``state_file`` exists, the pretrained model and</span>
-<span class="sd"> optimizer states are loaded.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> model : `pysegcnn.core.models.Network`</span>
-<span class="sd"> An instance of `pysegcnn.core.models.Network`.</span>
-<span class="sd"> optimizer : `torch.optim.Optimizer`</span>
-<span class="sd"> An instance of `torch.optim.Optimizer`.</span>
-<span class="sd"> state_file : `pathlib.Path`</span>
-<span class="sd"> Path to the model checkpoint.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> model : `pysegcnn.core.models.Network`</span>
-<span class="sd"> An instance of `pysegcnn.core.models.Network`.</span>
-<span class="sd"> optimizer : `torch.optim.Optimizer`</span>
-<span class="sd"> An instance of `torch.optim.Optimizer`.</span>
-<span class="sd"> checkpoint_state : `dict` [`str`, `numpy.ndarray`]</span>
-<span class="sd"> If the model checkpoint ``state_file`` exists, ``checkpoint_state``</span>
-<span class="sd"> has keys:</span>
-<span class="sd"> ``'ta'``</span>
-<span class="sd"> The accuracy on the training set (`numpy.ndarray`).</span>
-<span class="sd"> ``'tl'``</span>
-<span class="sd"> The loss on the training set (`numpy.ndarray`).</span>
-<span class="sd"> ``'va'``</span>
-<span class="sd"> The accuracy on the validation set (`numpy.ndarray`).</span>
-<span class="sd"> ``'vl'``</span>
-<span class="sd"> The loss on the validation set (`numpy.ndarray`).</span>
-
-<span class="sd"> """</span>
- <span class="c1"># whether to resume training from an existing model checkpoint</span>
- <span class="n">checkpoint_state</span> <span class="o">=</span> <span class="p">{}</span>
-
- <span class="c1"># if no checkpoint exists, file a warning and continue with a model</span>
- <span class="c1"># initialized from scratch</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="n">state_file</span><span class="o">.</span><span class="n">exists</span><span class="p">():</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">warning</span><span class="p">(</span><span class="s1">'Checkpoint for model </span><span class="si">{}</span><span class="s1"> does not exist. '</span>
- <span class="s1">'Initializing new model.'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">state_file</span><span class="o">.</span><span class="n">name</span><span class="p">))</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="c1"># load model checkpoint</span>
- <span class="n">model</span><span class="p">,</span> <span class="n">optimizer</span><span class="p">,</span> <span class="n">model_state</span> <span class="o">=</span> <span class="n">Network</span><span class="o">.</span><span class="n">load</span><span class="p">(</span><span class="n">state_file</span><span class="p">)</span>
-
- <span class="c1"># load model loss and accuracy</span>
-
- <span class="c1"># get all non-zero elements, i.e. get number of epochs trained</span>
- <span class="c1"># before the early stop</span>
- <span class="n">checkpoint_state</span> <span class="o">=</span> <span class="p">{</span><span class="n">k</span><span class="p">:</span> <span class="n">v</span><span class="p">[</span><span class="n">np</span><span class="o">.</span><span class="n">nonzero</span><span class="p">(</span><span class="n">v</span><span class="p">)]</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">v</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="o">-</span><span class="mi">1</span><span class="p">)</span>
- <span class="k">for</span> <span class="n">k</span><span class="p">,</span> <span class="n">v</span> <span class="ow">in</span> <span class="n">model_state</span><span class="p">[</span><span class="s1">'state'</span><span class="p">]</span><span class="o">.</span><span class="n">items</span><span class="p">()}</span>
-
- <span class="k">return</span> <span class="n">model</span><span class="p">,</span> <span class="n">optimizer</span><span class="p">,</span> <span class="n">checkpoint_state</span></div>
-
-<div class="viewcode-block" id="ModelConfig.transfer_model"><a class="viewcode-back" href="../../source/core.html#core.trainer.ModelConfig.transfer_model">[docs]</a> <span class="nd">@staticmethod</span>
- <span class="k">def</span> <span class="nf">transfer_model</span><span class="p">(</span><span class="n">state_file</span><span class="p">,</span> <span class="n">ds</span><span class="p">):</span>
- <span class="sd">"""Adjust a pretrained model to a new dataset.</span>
-
-<span class="sd"> The classification layer of the pretrained model in ``state_file`` is</span>
-<span class="sd"> initilialized from scratch with the classes of the new dataset ``ds``.</span>
-
-<span class="sd"> The remaining model weights are preserved.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> state_file : `pathlib.Path`</span>
-<span class="sd"> Path to a pretrained model.</span>
-<span class="sd"> ds : `pysegcnn.core.dataset.ImageDataset`</span>
-<span class="sd"> An instance of `pysegcnn.core.dataset.ImageDataset`.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> TypeError</span>
-<span class="sd"> Raised if ``ds`` is not an instance of</span>
-<span class="sd"> `pysegcnn.core.dataset.ImageDataset`.</span>
-<span class="sd"> ValueError</span>
-<span class="sd"> Raised if the bands of ``ds`` do not match the bands of the dataset</span>
-<span class="sd"> the pretrained model was trained with.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> model : `pysegcnn.core.models.Network`</span>
-<span class="sd"> An instance of `pysegcnn.core.models.Network`. The pretrained model</span>
-<span class="sd"> adjusted to the new dataset.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># check input type</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">ds</span><span class="p">,</span> <span class="n">ImageDataset</span><span class="p">):</span>
- <span class="k">raise</span> <span class="ne">TypeError</span><span class="p">(</span><span class="s1">'Expected "ds" to be </span><span class="si">{}</span><span class="s1">.'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="s1">'.'</span><span class="o">.</span><span class="n">join</span><span class="p">([</span><span class="n">ImageDataset</span><span class="o">.</span><span class="vm">__module__</span><span class="p">,</span>
- <span class="n">ImageDataset</span><span class="o">.</span><span class="vm">__name__</span><span class="p">])))</span>
-
- <span class="c1"># load the pretrained model</span>
- <span class="n">model</span><span class="p">,</span> <span class="n">_</span><span class="p">,</span> <span class="n">model_state</span> <span class="o">=</span> <span class="n">Network</span><span class="o">.</span><span class="n">load</span><span class="p">(</span><span class="n">state_file</span><span class="p">)</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Configuring model for new dataset: </span><span class="si">{}</span><span class="s1">.'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span>
- <span class="n">ds</span><span class="o">.</span><span class="vm">__class__</span><span class="o">.</span><span class="vm">__name__</span><span class="p">))</span>
-
- <span class="c1"># check whether the current dataset uses the correct spectral bands</span>
- <span class="k">if</span> <span class="n">ds</span><span class="o">.</span><span class="n">use_bands</span> <span class="o">!=</span> <span class="n">model_state</span><span class="p">[</span><span class="s1">'bands'</span><span class="p">]:</span>
- <span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="s1">'The pretrained network was trained with '</span>
- <span class="s1">'bands </span><span class="si">{}</span><span class="s1">, not with bands </span><span class="si">{}</span><span class="s1">.'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">model_state</span><span class="p">[</span><span class="s1">'bands'</span><span class="p">],</span> <span class="n">ds</span><span class="o">.</span><span class="n">use_bands</span><span class="p">))</span>
-
- <span class="c1"># get the number of convolutional filters</span>
- <span class="n">filters</span> <span class="o">=</span> <span class="n">model_state</span><span class="p">[</span><span class="s1">'params'</span><span class="p">][</span><span class="s1">'filters'</span><span class="p">]</span>
-
- <span class="c1"># reset model epoch to 0, since the model is trained on a different</span>
- <span class="c1"># dataset</span>
- <span class="n">model</span><span class="o">.</span><span class="n">epoch</span> <span class="o">=</span> <span class="mi">0</span>
-
- <span class="c1"># adjust the number of classes in the model</span>
- <span class="n">model</span><span class="o">.</span><span class="n">nclasses</span> <span class="o">=</span> <span class="nb">len</span><span class="p">(</span><span class="n">ds</span><span class="o">.</span><span class="n">labels</span><span class="p">)</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Replacing classification layer to classes: </span><span class="si">{}</span><span class="s1">.'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="s1">', '</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="s1">'(</span><span class="si">{}</span><span class="s1">, </span><span class="si">{}</span><span class="s1">)'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">k</span><span class="p">,</span> <span class="n">v</span><span class="p">[</span><span class="s1">'label'</span><span class="p">])</span>
- <span class="k">for</span> <span class="n">k</span><span class="p">,</span> <span class="n">v</span> <span class="ow">in</span> <span class="n">ds</span><span class="o">.</span><span class="n">labels</span><span class="o">.</span><span class="n">items</span><span class="p">())))</span>
-
- <span class="c1"># adjust the classification layer to the classes of the new dataset</span>
- <span class="n">model</span><span class="o">.</span><span class="n">classifier</span> <span class="o">=</span> <span class="n">Conv2dSame</span><span class="p">(</span><span class="n">in_channels</span><span class="o">=</span><span class="n">filters</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span>
- <span class="n">out_channels</span><span class="o">=</span><span class="n">model</span><span class="o">.</span><span class="n">nclasses</span><span class="p">,</span>
- <span class="n">kernel_size</span><span class="o">=</span><span class="mi">1</span><span class="p">)</span>
-
- <span class="k">return</span> <span class="n">model</span></div></div>
-
-
-<div class="viewcode-block" id="StateConfig"><a class="viewcode-back" href="../../source/core.html#core.trainer.StateConfig">[docs]</a><span class="nd">@dataclasses</span><span class="o">.</span><span class="n">dataclass</span>
-<span class="k">class</span> <span class="nc">StateConfig</span><span class="p">(</span><span class="n">BaseConfig</span><span class="p">):</span>
- <span class="sd">"""Model state configuration class.</span>
-
-<span class="sd"> Generate the model state filename according to the following naming</span>
-<span class="sd"> convention:</span>
-
-<span class="sd"> model_dataset_optimizer_splitmode_splitparams_tilesize_batchsize_bands.pt</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> ds : `pysegcnn.core.dataset.ImageDataset`</span>
-<span class="sd"> An instance of `pysegcnn.core.dataset.ImageDataset`.</span>
-<span class="sd"> sc : `pysegcnn.core.trainer.SplitConfig`</span>
-<span class="sd"> An instance of `pysegcnn.core.trainer.SplitConfig`.</span>
-<span class="sd"> mc : `pysegcnn.core.trainer.ModelConfig`</span>
-<span class="sd"> An instance of `pysegcnn.core.trainer.SplitConfig`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="n">ds</span><span class="p">:</span> <span class="n">ImageDataset</span>
- <span class="n">sc</span><span class="p">:</span> <span class="n">SplitConfig</span>
- <span class="n">mc</span><span class="p">:</span> <span class="n">ModelConfig</span>
-
- <span class="k">def</span> <span class="nf">__post_init__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Check the type of each argument.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="n">__post_init__</span><span class="p">()</span>
-
-<div class="viewcode-block" id="StateConfig.init_state"><a class="viewcode-back" href="../../source/core.html#core.trainer.StateConfig.init_state">[docs]</a> <span class="k">def</span> <span class="nf">init_state</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Generate the model state filename.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> state : `pathlib.Path`</span>
-<span class="sd"> The path to the model state file.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># file to save model state to:</span>
- <span class="c1"># network_dataset_optim_split_splitparams_tilesize_batchsize_bands.pt</span>
-
- <span class="c1"># model state filename</span>
- <span class="n">state_file</span> <span class="o">=</span> <span class="s1">'</span><span class="si">{}</span><span class="s1">_</span><span class="si">{}</span><span class="s1">_</span><span class="si">{}</span><span class="s1">_</span><span class="si">{}</span><span class="s1">Split_</span><span class="si">{}</span><span class="s1">_t</span><span class="si">{}</span><span class="s1">_b</span><span class="si">{}</span><span class="s1">_</span><span class="si">{}</span><span class="s1">.pt'</span>
-
- <span class="c1"># get the band numbers</span>
- <span class="n">bformat</span> <span class="o">=</span> <span class="s1">''</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">band</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">+</span>
- <span class="nb">str</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">ds</span><span class="o">.</span><span class="n">sensor</span><span class="o">.</span><span class="n">__members__</span><span class="p">[</span><span class="n">band</span><span class="p">]</span><span class="o">.</span><span class="n">value</span><span class="p">)</span> <span class="k">for</span>
- <span class="n">band</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">ds</span><span class="o">.</span><span class="n">use_bands</span><span class="p">)</span>
-
- <span class="c1"># check which split mode was used</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">sc</span><span class="o">.</span><span class="n">split_mode</span> <span class="o">==</span> <span class="s1">'date'</span><span class="p">:</span>
- <span class="c1"># store the date that was used to split the dataset</span>
- <span class="n">state_file</span> <span class="o">=</span> <span class="n">state_file</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">mc</span><span class="o">.</span><span class="n">model_name</span><span class="p">,</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">ds</span><span class="o">.</span><span class="vm">__class__</span><span class="o">.</span><span class="vm">__name__</span><span class="p">,</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">mc</span><span class="o">.</span><span class="n">optim_name</span><span class="p">,</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">sc</span><span class="o">.</span><span class="n">split_mode</span><span class="o">.</span><span class="n">capitalize</span><span class="p">(),</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">sc</span><span class="o">.</span><span class="n">date</span><span class="p">,</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">ds</span><span class="o">.</span><span class="n">tile_size</span><span class="p">,</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">mc</span><span class="o">.</span><span class="n">batch_size</span><span class="p">,</span>
- <span class="n">bformat</span><span class="p">)</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="c1"># store the random split parameters</span>
- <span class="n">split_params</span> <span class="o">=</span> <span class="s1">'s</span><span class="si">{}</span><span class="s1">_t</span><span class="si">{}</span><span class="s1">v</span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">ds</span><span class="o">.</span><span class="n">seed</span><span class="p">,</span> <span class="nb">str</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">sc</span><span class="o">.</span><span class="n">ttratio</span><span class="p">)</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="s1">'.'</span><span class="p">,</span> <span class="s1">''</span><span class="p">),</span>
- <span class="nb">str</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">sc</span><span class="o">.</span><span class="n">tvratio</span><span class="p">)</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="s1">'.'</span><span class="p">,</span> <span class="s1">''</span><span class="p">))</span>
-
- <span class="c1"># model state filename</span>
- <span class="n">state_file</span> <span class="o">=</span> <span class="n">state_file</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">mc</span><span class="o">.</span><span class="n">model_name</span><span class="p">,</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">ds</span><span class="o">.</span><span class="vm">__class__</span><span class="o">.</span><span class="vm">__name__</span><span class="p">,</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">mc</span><span class="o">.</span><span class="n">optim_name</span><span class="p">,</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">sc</span><span class="o">.</span><span class="n">split_mode</span><span class="o">.</span><span class="n">capitalize</span><span class="p">(),</span>
- <span class="n">split_params</span><span class="p">,</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">ds</span><span class="o">.</span><span class="n">tile_size</span><span class="p">,</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">mc</span><span class="o">.</span><span class="n">batch_size</span><span class="p">,</span>
- <span class="n">bformat</span><span class="p">)</span>
-
- <span class="c1"># check whether a pretrained model was used and change state filename</span>
- <span class="c1"># accordingly</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">mc</span><span class="o">.</span><span class="n">transfer</span><span class="p">:</span>
- <span class="c1"># add the configuration of the pretrained model to the state name</span>
- <span class="n">state_file</span> <span class="o">=</span> <span class="p">(</span><span class="n">state_file</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="s1">'.pt'</span><span class="p">,</span> <span class="s1">'_'</span><span class="p">)</span> <span class="o">+</span>
- <span class="s1">'pretrained_'</span> <span class="o">+</span> <span class="bp">self</span><span class="o">.</span><span class="n">mc</span><span class="o">.</span><span class="n">pretrained_model</span><span class="p">)</span>
-
- <span class="c1"># path to model state</span>
- <span class="n">state</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">mc</span><span class="o">.</span><span class="n">state_path</span><span class="o">.</span><span class="n">joinpath</span><span class="p">(</span><span class="n">state_file</span><span class="p">)</span>
-
- <span class="k">return</span> <span class="n">state</span></div></div>
-
-
-<div class="viewcode-block" id="EvalConfig"><a class="viewcode-back" href="../../source/core.html#core.trainer.EvalConfig">[docs]</a><span class="nd">@dataclasses</span><span class="o">.</span><span class="n">dataclass</span>
-<span class="k">class</span> <span class="nc">EvalConfig</span><span class="p">(</span><span class="n">BaseConfig</span><span class="p">):</span>
- <span class="sd">"""Model inference configuration.</span>
-
-<span class="sd"> Evaluate a model.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> state_file : `pathlib.Path`</span>
-<span class="sd"> Path to the model to evaluate.</span>
-<span class="sd"> test : `bool` or `None`</span>
-<span class="sd"> Whether to evaluate the model on the training(``test`` = `None`), the</span>
-<span class="sd"> validation (``test`` = False) or the test set (``test`` = True).</span>
-<span class="sd"> predict_scene : `bool`, optional</span>
-<span class="sd"> The model prediction order. If False, the samples (tiles) of a dataset</span>
-<span class="sd"> are predicted in any order and the scenes are not reconstructed.</span>
-<span class="sd"> If True, the samples (tiles) are ordered according to the scene they</span>
-<span class="sd"> belong to and a model prediction for each entire reconstructed scene is</span>
-<span class="sd"> returned. The default is False.</span>
-<span class="sd"> plot_samples : `bool`, optional</span>
-<span class="sd"> Whether to save a plot of false color composite, ground truth and model</span>
-<span class="sd"> prediction for each sample (tile). Only used if ``predict_scene`` =</span>
-<span class="sd"> False. The default is False.</span>
-<span class="sd"> plot_scenes : `bool`, optional</span>
-<span class="sd"> Whether to save a plot of false color composite, ground truth and model</span>
-<span class="sd"> prediction for each entire scene. Only used if ``predict_scene`` =</span>
-<span class="sd"> True. The default is False.</span>
-<span class="sd"> plot_bands : `list` [`str`], optional</span>
-<span class="sd"> The bands to build the false color composite. The default is</span>
-<span class="sd"> ['nir', 'red', 'green'].</span>
-<span class="sd"> cm : `bool`, optional</span>
-<span class="sd"> Whether to compute and plot the confusion matrix. The default is True.</span>
-<span class="sd"> figsize : `tuple`, optional</span>
-<span class="sd"> The figure size in centimeters. The default is (10, 10).</span>
-<span class="sd"> alpha : `int`, optional</span>
-<span class="sd"> The level of the percentiles for contrast stretching of the false color</span>
-<span class="sd"> compsite. The default is 0, i.e. no stretching.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="n">state_file</span><span class="p">:</span> <span class="n">pathlib</span><span class="o">.</span><span class="n">Path</span>
- <span class="n">test</span><span class="p">:</span> <span class="nb">object</span>
- <span class="n">predict_scene</span><span class="p">:</span> <span class="nb">bool</span> <span class="o">=</span> <span class="kc">False</span>
- <span class="n">plot_samples</span><span class="p">:</span> <span class="nb">bool</span> <span class="o">=</span> <span class="kc">False</span>
- <span class="n">plot_scenes</span><span class="p">:</span> <span class="nb">bool</span> <span class="o">=</span> <span class="kc">False</span>
- <span class="n">plot_bands</span><span class="p">:</span> <span class="nb">list</span> <span class="o">=</span> <span class="n">dataclasses</span><span class="o">.</span><span class="n">field</span><span class="p">(</span>
- <span class="n">default_factory</span><span class="o">=</span><span class="k">lambda</span><span class="p">:</span> <span class="p">[</span><span class="s1">'nir'</span><span class="p">,</span> <span class="s1">'red'</span><span class="p">,</span> <span class="s1">'green'</span><span class="p">])</span>
- <span class="n">cm</span><span class="p">:</span> <span class="nb">bool</span> <span class="o">=</span> <span class="kc">True</span>
- <span class="n">figsize</span><span class="p">:</span> <span class="nb">tuple</span> <span class="o">=</span> <span class="p">(</span><span class="mi">10</span><span class="p">,</span> <span class="mi">10</span><span class="p">)</span>
- <span class="n">alpha</span><span class="p">:</span> <span class="nb">int</span> <span class="o">=</span> <span class="mi">5</span>
-
- <span class="k">def</span> <span class="nf">__post_init__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Check the type of each argument.</span>
-
-<span class="sd"> Configure figure output paths.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> TypeError</span>
-<span class="sd"> Raised if ``test`` is not of type `bool` or `None`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="n">__post_init__</span><span class="p">()</span>
-
- <span class="c1"># check whether the test input parameter is correctly specified</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">test</span> <span class="ow">not</span> <span class="ow">in</span> <span class="p">[</span><span class="kc">None</span><span class="p">,</span> <span class="kc">False</span><span class="p">,</span> <span class="kc">True</span><span class="p">]:</span>
- <span class="k">raise</span> <span class="ne">TypeError</span><span class="p">(</span><span class="s1">'Expected "test" to be None, True or False, got '</span>
- <span class="s1">'</span><span class="si">{}</span><span class="s1">.'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">test</span><span class="p">))</span>
-
- <span class="c1"># the output paths for the different graphics</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">base_path</span> <span class="o">=</span> <span class="n">pathlib</span><span class="o">.</span><span class="n">Path</span><span class="p">(</span><span class="n">HERE</span><span class="p">)</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">sample_path</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">base_path</span><span class="o">.</span><span class="n">joinpath</span><span class="p">(</span><span class="s1">'_samples'</span><span class="p">)</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">scenes_path</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">base_path</span><span class="o">.</span><span class="n">joinpath</span><span class="p">(</span><span class="s1">'_scenes'</span><span class="p">)</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">perfmc_path</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">base_path</span><span class="o">.</span><span class="n">joinpath</span><span class="p">(</span><span class="s1">'_graphics'</span><span class="p">)</span>
-
- <span class="c1"># input path for model state files</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">models_path</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">base_path</span><span class="o">.</span><span class="n">joinpath</span><span class="p">(</span><span class="s1">'_models'</span><span class="p">)</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">state_file</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">models_path</span><span class="o">.</span><span class="n">joinpath</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">state_file</span><span class="p">)</span>
-
- <span class="c1"># write initialization string to log file</span>
- <span class="n">LogConfig</span><span class="o">.</span><span class="n">init_log</span><span class="p">(</span><span class="s1">'</span><span class="si">{}</span><span class="s1">: '</span> <span class="o">+</span> <span class="s1">'Evaluating model: </span><span class="si">{}</span><span class="s1">.'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">state_file</span><span class="o">.</span><span class="n">name</span><span class="p">))</span></div>
-
-
-<div class="viewcode-block" id="LogConfig"><a class="viewcode-back" href="../../source/core.html#core.trainer.LogConfig">[docs]</a><span class="nd">@dataclasses</span><span class="o">.</span><span class="n">dataclass</span>
-<span class="k">class</span> <span class="nc">LogConfig</span><span class="p">(</span><span class="n">BaseConfig</span><span class="p">):</span>
- <span class="sd">"""Logging configuration class.</span>
-
-<span class="sd"> Generate the model log file.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> state_file : `pathlib.Path`</span>
-<span class="sd"> Path to a model state file.</span>
-
-<span class="sd"> """</span>
-
- <span class="n">state_file</span><span class="p">:</span> <span class="n">pathlib</span><span class="o">.</span><span class="n">Path</span>
-
- <span class="k">def</span> <span class="nf">__post_init__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Check the type of each argument.</span>
-
-<span class="sd"> Generate model log file.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="n">__post_init__</span><span class="p">()</span>
-
- <span class="c1"># the path to store model logs</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">log_path</span> <span class="o">=</span> <span class="n">pathlib</span><span class="o">.</span><span class="n">Path</span><span class="p">(</span><span class="n">HERE</span><span class="p">)</span><span class="o">.</span><span class="n">joinpath</span><span class="p">(</span><span class="s1">'_logs'</span><span class="p">)</span>
-
- <span class="c1"># the log file of the current model</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">log_file</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">log_path</span><span class="o">.</span><span class="n">joinpath</span><span class="p">(</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">state_file</span><span class="o">.</span><span class="n">name</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="s1">'.pt'</span><span class="p">,</span> <span class="s1">'.log'</span><span class="p">))</span>
-
-<div class="viewcode-block" id="LogConfig.now"><a class="viewcode-back" href="../../source/core.html#core.trainer.LogConfig.now">[docs]</a> <span class="nd">@staticmethod</span>
- <span class="k">def</span> <span class="nf">now</span><span class="p">():</span>
- <span class="sd">"""Return the current date and time.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> date : `datetime.datetime`</span>
-<span class="sd"> The current date and time.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="n">datetime</span><span class="o">.</span><span class="n">datetime</span><span class="o">.</span><span class="n">strftime</span><span class="p">(</span><span class="n">datetime</span><span class="o">.</span><span class="n">datetime</span><span class="o">.</span><span class="n">now</span><span class="p">(),</span>
- <span class="s1">'%Y-%m-</span><span class="si">%d</span><span class="s1">T%H:%M:%S'</span><span class="p">)</span></div>
-
-<div class="viewcode-block" id="LogConfig.init_log"><a class="viewcode-back" href="../../source/core.html#core.trainer.LogConfig.init_log">[docs]</a> <span class="nd">@staticmethod</span>
- <span class="k">def</span> <span class="nf">init_log</span><span class="p">(</span><span class="n">init_str</span><span class="p">):</span>
- <span class="sd">"""Generate a string to identify a new model run.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> init_str : `str`</span>
-<span class="sd"> The string to write to the model log file.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="mi">80</span> <span class="o">*</span> <span class="s1">'-'</span><span class="p">)</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="n">init_str</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">LogConfig</span><span class="o">.</span><span class="n">now</span><span class="p">()))</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="mi">80</span> <span class="o">*</span> <span class="s1">'-'</span><span class="p">)</span></div></div>
-
-
-<div class="viewcode-block" id="NetworkTrainer"><a class="viewcode-back" href="../../source/core.html#core.trainer.NetworkTrainer">[docs]</a><span class="nd">@dataclasses</span><span class="o">.</span><span class="n">dataclass</span>
-<span class="k">class</span> <span class="nc">NetworkTrainer</span><span class="p">(</span><span class="n">BaseConfig</span><span class="p">):</span>
- <span class="sd">"""Model training class.</span>
-
-<span class="sd"> Generic class to train an instance of `pysegcnn.core.models.Network` on</span>
-<span class="sd"> a dataset of type `pysegcnn.core.dataset.ImageDataset`.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> model : `pysegcnn.core.models.Network`</span>
-<span class="sd"> The model to train. An instance of `pysegcnn.core.models.Network`.</span>
-<span class="sd"> optimizer : `torch.optim.Optimizer`</span>
-<span class="sd"> The optimizer to update the model weights. An instance of</span>
-<span class="sd"> `torch.optim.Optimizer`.</span>
-<span class="sd"> loss_function : `torch.nn.Module`</span>
-<span class="sd"> The loss function to compute the model error. An instance of</span>
-<span class="sd"> `torch.nn.Module`.</span>
-<span class="sd"> train_dl : `torch.utils.data.DataLoader`</span>
-<span class="sd"> The training `torch.utils.data.DataLoader` instance.</span>
-<span class="sd"> valid_dl : `torch.utils.data.DataLoader`</span>
-<span class="sd"> The validation `torch.utils.data.DataLoader` instance.</span>
-<span class="sd"> test_dl : `torch.utils.data.DataLoader`</span>
-<span class="sd"> The test `torch.utils.data.DataLoader` instance.</span>
-<span class="sd"> state_file : `pathlib.Path`</span>
-<span class="sd"> Path to save the model state.</span>
-<span class="sd"> epochs : `int`, optional</span>
-<span class="sd"> The maximum number of epochs to train. The default is 1.</span>
-<span class="sd"> nthreads : `int`, optional</span>
-<span class="sd"> The number of cpu threads to use during training. The default is</span>
-<span class="sd"> torch.get_num_threads().</span>
-<span class="sd"> early_stop : `bool`, optional</span>
-<span class="sd"> Whether to apply `early stopping`_. The default is False.</span>
-<span class="sd"> mode : `str`, optional</span>
-<span class="sd"> The mode of the early stopping. Depends on the metric measuring</span>
-<span class="sd"> performance. When using model loss as metric, use ``mode`` = 'min',</span>
-<span class="sd"> however, when using accuracy as metric, use ``mode`` = 'max'. For now,</span>
-<span class="sd"> only ``mode`` = 'max' is supported. Only used if ``early_stop`` = True.</span>
-<span class="sd"> The default is 'max'.</span>
-<span class="sd"> delta : `float`, optional</span>
-<span class="sd"> Minimum change in early stopping metric to be considered as an</span>
-<span class="sd"> improvement. Only used if ``early_stop`` = True. The default is 0.</span>
-<span class="sd"> patience : `int`, optional</span>
-<span class="sd"> The number of epochs to wait for an improvement in the early stopping</span>
-<span class="sd"> metric. If the model does not improve over more than ``patience``</span>
-<span class="sd"> epochs, quit training. Only used if ``early_stop`` = True.</span>
-<span class="sd"> The default is 10.</span>
-<span class="sd"> checkpoint_state : `dict` [`str`, `numpy.ndarray`], optional</span>
-<span class="sd"> A model checkpoint for ``model``. If specified, ``checkpoint_state``</span>
-<span class="sd"> should be a dictionary with keys:</span>
-<span class="sd"> ``'ta'``</span>
-<span class="sd"> The accuracy on the training set (`numpy.ndarray`).</span>
-<span class="sd"> ``'tl'``</span>
-<span class="sd"> The loss on the training set (`numpy.ndarray`).</span>
-<span class="sd"> ``'va'``</span>
-<span class="sd"> The accuracy on the validation set (`numpy.ndarray`).</span>
-<span class="sd"> ``'vl'``</span>
-<span class="sd"> The loss on the validation set (`numpy.ndarray`).</span>
-<span class="sd"> The default is {}.</span>
-<span class="sd"> save : `bool`, optional</span>
-<span class="sd"> Whether to save the model state to ``state_file``. The default is True.</span>
-
-<span class="sd"> .. _early stopping:</span>
-<span class="sd"> https://en.wikipedia.org/wiki/Early_stopping</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="n">model</span><span class="p">:</span> <span class="n">Network</span>
- <span class="n">optimizer</span><span class="p">:</span> <span class="n">Optimizer</span>
- <span class="n">loss_function</span><span class="p">:</span> <span class="n">nn</span><span class="o">.</span><span class="n">Module</span>
- <span class="n">train_dl</span><span class="p">:</span> <span class="n">DataLoader</span>
- <span class="n">valid_dl</span><span class="p">:</span> <span class="n">DataLoader</span>
- <span class="n">test_dl</span><span class="p">:</span> <span class="n">DataLoader</span>
- <span class="n">state_file</span><span class="p">:</span> <span class="n">pathlib</span><span class="o">.</span><span class="n">Path</span>
- <span class="n">epochs</span><span class="p">:</span> <span class="nb">int</span> <span class="o">=</span> <span class="mi">1</span>
- <span class="n">nthreads</span><span class="p">:</span> <span class="nb">int</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">get_num_threads</span><span class="p">()</span>
- <span class="n">early_stop</span><span class="p">:</span> <span class="nb">bool</span> <span class="o">=</span> <span class="kc">False</span>
- <span class="n">mode</span><span class="p">:</span> <span class="nb">str</span> <span class="o">=</span> <span class="s1">'max'</span>
- <span class="n">delta</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mi">0</span>
- <span class="n">patience</span><span class="p">:</span> <span class="nb">int</span> <span class="o">=</span> <span class="mi">10</span>
- <span class="n">checkpoint_state</span><span class="p">:</span> <span class="nb">dict</span> <span class="o">=</span> <span class="n">dataclasses</span><span class="o">.</span><span class="n">field</span><span class="p">(</span><span class="n">default_factory</span><span class="o">=</span><span class="nb">dict</span><span class="p">)</span>
- <span class="n">save</span><span class="p">:</span> <span class="nb">bool</span> <span class="o">=</span> <span class="kc">True</span>
-
- <span class="k">def</span> <span class="nf">__post_init__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Check the type of each argument.</span>
-
-<span class="sd"> Configure the device to train the model on, i.e. train on the gpu if</span>
-<span class="sd"> available.</span>
-
-<span class="sd"> Configure early stopping if required.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="n">__post_init__</span><span class="p">()</span>
-
- <span class="c1"># whether to use the gpu</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">device</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">device</span><span class="p">(</span><span class="s2">"cuda:0"</span> <span class="k">if</span> <span class="n">torch</span><span class="o">.</span><span class="n">cuda</span><span class="o">.</span><span class="n">is_available</span><span class="p">()</span>
- <span class="k">else</span> <span class="s2">"cpu"</span><span class="p">)</span>
-
- <span class="c1"># send the model to the gpu if available</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">model</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">model</span><span class="o">.</span><span class="n">to</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">device</span><span class="p">)</span>
-
- <span class="c1"># maximum accuracy on the validation dataset</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">max_accuracy</span> <span class="o">=</span> <span class="mi">0</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">checkpoint_state</span><span class="p">:</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">max_accuracy</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">checkpoint_state</span><span class="p">[</span><span class="s1">'va'</span><span class="p">]</span><span class="o">.</span><span class="n">mean</span><span class="p">(</span>
- <span class="n">axis</span><span class="o">=</span><span class="mi">0</span><span class="p">)</span><span class="o">.</span><span class="n">max</span><span class="p">()</span><span class="o">.</span><span class="n">item</span><span class="p">()</span>
-
- <span class="c1"># whether to use early stopping</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">es</span> <span class="o">=</span> <span class="kc">None</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">early_stop</span><span class="p">:</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">es</span> <span class="o">=</span> <span class="n">EarlyStopping</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">mode</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">max_accuracy</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">delta</span><span class="p">,</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">patience</span><span class="p">)</span>
-
- <span class="c1"># log representation</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="nb">repr</span><span class="p">(</span><span class="bp">self</span><span class="p">))</span>
-
-<div class="viewcode-block" id="NetworkTrainer.train"><a class="viewcode-back" href="../../source/core.html#core.trainer.NetworkTrainer.train">[docs]</a> <span class="k">def</span> <span class="nf">train</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Train the model.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> training_state : `dict` [`str`, `numpy.ndarray`]</span>
-<span class="sd"> The training state dictionary with keys:</span>
-<span class="sd"> ``'ta'``</span>
-<span class="sd"> The accuracy on the training set (`numpy.ndarray`).</span>
-<span class="sd"> ``'tl'``</span>
-<span class="sd"> The loss on the training set (`numpy.ndarray`).</span>
-<span class="sd"> ``'va'``</span>
-<span class="sd"> The accuracy on the validation set (`numpy.ndarray`).</span>
-<span class="sd"> ``'vl'``</span>
-<span class="sd"> The loss on the validation set (`numpy.ndarray`).</span>
-
-<span class="sd"> """</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="mi">35</span> <span class="o">*</span> <span class="s1">'-'</span> <span class="o">+</span> <span class="s1">' Training '</span> <span class="o">+</span> <span class="mi">35</span> <span class="o">*</span> <span class="s1">'-'</span><span class="p">)</span>
-
- <span class="c1"># set the number of threads</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Device: </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">device</span><span class="p">))</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Number of cpu threads: </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">nthreads</span><span class="p">))</span>
- <span class="n">torch</span><span class="o">.</span><span class="n">set_num_threads</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">nthreads</span><span class="p">)</span>
-
- <span class="c1"># create dictionary of the observed losses and accuracies on the</span>
- <span class="c1"># training and validation dataset</span>
- <span class="n">tshape</span> <span class="o">=</span> <span class="p">(</span><span class="nb">len</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">train_dl</span><span class="p">),</span> <span class="bp">self</span><span class="o">.</span><span class="n">epochs</span><span class="p">)</span>
- <span class="n">vshape</span> <span class="o">=</span> <span class="p">(</span><span class="nb">len</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">valid_dl</span><span class="p">),</span> <span class="bp">self</span><span class="o">.</span><span class="n">epochs</span><span class="p">)</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">training_state</span> <span class="o">=</span> <span class="p">{</span><span class="s1">'tl'</span><span class="p">:</span> <span class="n">np</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="n">tshape</span><span class="p">),</span>
- <span class="s1">'ta'</span><span class="p">:</span> <span class="n">np</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="n">tshape</span><span class="p">),</span>
- <span class="s1">'vl'</span><span class="p">:</span> <span class="n">np</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="n">vshape</span><span class="p">),</span>
- <span class="s1">'va'</span><span class="p">:</span> <span class="n">np</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="n">vshape</span><span class="p">)</span>
- <span class="p">}</span>
-
- <span class="c1"># initialize the training: iterate over the entire training dataset</span>
- <span class="k">for</span> <span class="n">epoch</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">epochs</span><span class="p">):</span>
-
- <span class="c1"># set the model to training mode</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Setting model to training mode ...'</span><span class="p">)</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">model</span><span class="o">.</span><span class="n">train</span><span class="p">()</span>
-
- <span class="c1"># iterate over the dataloader object</span>
- <span class="k">for</span> <span class="n">batch</span><span class="p">,</span> <span class="p">(</span><span class="n">inputs</span><span class="p">,</span> <span class="n">labels</span><span class="p">)</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">train_dl</span><span class="p">):</span>
-
- <span class="c1"># send the data to the gpu if available</span>
- <span class="n">inputs</span> <span class="o">=</span> <span class="n">inputs</span><span class="o">.</span><span class="n">to</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">device</span><span class="p">)</span>
- <span class="n">labels</span> <span class="o">=</span> <span class="n">labels</span><span class="o">.</span><span class="n">to</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">device</span><span class="p">)</span>
-
- <span class="c1"># reset the gradients</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">optimizer</span><span class="o">.</span><span class="n">zero_grad</span><span class="p">()</span>
-
- <span class="c1"># perform forward pass</span>
- <span class="n">outputs</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">model</span><span class="p">(</span><span class="n">inputs</span><span class="p">)</span>
-
- <span class="c1"># compute loss</span>
- <span class="n">loss</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">loss_function</span><span class="p">(</span><span class="n">outputs</span><span class="p">,</span> <span class="n">labels</span><span class="o">.</span><span class="n">long</span><span class="p">())</span>
- <span class="n">observed_loss</span> <span class="o">=</span> <span class="n">loss</span><span class="o">.</span><span class="n">detach</span><span class="p">()</span><span class="o">.</span><span class="n">numpy</span><span class="p">()</span><span class="o">.</span><span class="n">item</span><span class="p">()</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">training_state</span><span class="p">[</span><span class="s1">'tl'</span><span class="p">][</span><span class="n">batch</span><span class="p">,</span> <span class="n">epoch</span><span class="p">]</span> <span class="o">=</span> <span class="n">observed_loss</span>
-
- <span class="c1"># compute the gradients of the loss function w.r.t.</span>
- <span class="c1"># the network weights</span>
- <span class="n">loss</span><span class="o">.</span><span class="n">backward</span><span class="p">()</span>
-
- <span class="c1"># update the weights</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">optimizer</span><span class="o">.</span><span class="n">step</span><span class="p">()</span>
-
- <span class="c1"># calculate predicted class labels</span>
- <span class="n">ypred</span> <span class="o">=</span> <span class="n">F</span><span class="o">.</span><span class="n">softmax</span><span class="p">(</span><span class="n">outputs</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="mi">1</span><span class="p">)</span><span class="o">.</span><span class="n">argmax</span><span class="p">(</span><span class="n">dim</span><span class="o">=</span><span class="mi">1</span><span class="p">)</span>
-
- <span class="c1"># calculate accuracy on current batch</span>
- <span class="n">observed_accuracy</span> <span class="o">=</span> <span class="n">accuracy_function</span><span class="p">(</span><span class="n">ypred</span><span class="p">,</span> <span class="n">labels</span><span class="p">)</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">training_state</span><span class="p">[</span><span class="s1">'ta'</span><span class="p">][</span><span class="n">batch</span><span class="p">,</span> <span class="n">epoch</span><span class="p">]</span> <span class="o">=</span> <span class="n">observed_accuracy</span>
-
- <span class="c1"># print progress</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Epoch: </span><span class="si">{:d}</span><span class="s1">/</span><span class="si">{:d}</span><span class="s1">, Mini-batch: </span><span class="si">{:d}</span><span class="s1">/</span><span class="si">{:d}</span><span class="s1">, '</span>
- <span class="s1">'Loss: </span><span class="si">{:.2f}</span><span class="s1">, Accuracy: </span><span class="si">{:.2f}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span>
- <span class="n">epoch</span> <span class="o">+</span> <span class="mi">1</span><span class="p">,</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">epochs</span><span class="p">,</span>
- <span class="n">batch</span> <span class="o">+</span> <span class="mi">1</span><span class="p">,</span>
- <span class="nb">len</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">train_dl</span><span class="p">),</span>
- <span class="n">observed_loss</span><span class="p">,</span>
- <span class="n">observed_accuracy</span><span class="p">))</span>
-
- <span class="c1"># update the number of epochs trained</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">model</span><span class="o">.</span><span class="n">epoch</span> <span class="o">+=</span> <span class="mi">1</span>
-
- <span class="c1"># whether to evaluate model performance on the validation set and</span>
- <span class="c1"># early stop the training process</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">early_stop</span><span class="p">:</span>
-
- <span class="c1"># model predictions on the validation set</span>
- <span class="n">vacc</span><span class="p">,</span> <span class="n">vloss</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">predict</span><span class="p">()</span>
-
- <span class="c1"># append observed accuracy and loss to arrays</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">training_state</span><span class="p">[</span><span class="s1">'va'</span><span class="p">][:,</span> <span class="n">epoch</span><span class="p">]</span> <span class="o">=</span> <span class="n">vacc</span><span class="o">.</span><span class="n">squeeze</span><span class="p">()</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">training_state</span><span class="p">[</span><span class="s1">'vl'</span><span class="p">][:,</span> <span class="n">epoch</span><span class="p">]</span> <span class="o">=</span> <span class="n">vloss</span><span class="o">.</span><span class="n">squeeze</span><span class="p">()</span>
-
- <span class="c1"># metric to assess model performance on the validation set</span>
- <span class="n">epoch_acc</span> <span class="o">=</span> <span class="n">vacc</span><span class="o">.</span><span class="n">squeeze</span><span class="p">()</span><span class="o">.</span><span class="n">mean</span><span class="p">()</span>
-
- <span class="c1"># whether the model improved with respect to the previous epoch</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">es</span><span class="o">.</span><span class="n">increased</span><span class="p">(</span><span class="n">epoch_acc</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">max_accuracy</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">delta</span><span class="p">):</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">max_accuracy</span> <span class="o">=</span> <span class="n">epoch_acc</span>
-
- <span class="c1"># save model state if the model improved with</span>
- <span class="c1"># respect to the previous epoch</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">save_state</span><span class="p">()</span>
-
- <span class="c1"># whether the early stopping criterion is met</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">es</span><span class="o">.</span><span class="n">stop</span><span class="p">(</span><span class="n">epoch_acc</span><span class="p">):</span>
- <span class="k">break</span>
-
- <span class="k">else</span><span class="p">:</span>
- <span class="c1"># if no early stopping is required, the model state is</span>
- <span class="c1"># saved after each epoch</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">save_state</span><span class="p">()</span>
-
- <span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">training_state</span></div>
-
-<div class="viewcode-block" id="NetworkTrainer.predict"><a class="viewcode-back" href="../../source/core.html#core.trainer.NetworkTrainer.predict">[docs]</a> <span class="k">def</span> <span class="nf">predict</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Model inference at training time.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> accuracies : `numpy.ndarray`</span>
-<span class="sd"> The mean model prediction accuracy on each mini-batch in the</span>
-<span class="sd"> validation set.</span>
-<span class="sd"> losses : `numpy.ndarray`</span>
-<span class="sd"> The model loss for each mini-batch in the validation set.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># set the model to evaluation mode</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Setting model to evaluation mode ...'</span><span class="p">)</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">model</span><span class="o">.</span><span class="n">eval</span><span class="p">()</span>
-
- <span class="c1"># create arrays of the observed losses and accuracies</span>
- <span class="n">accuracies</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="p">(</span><span class="nb">len</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">valid_dl</span><span class="p">),</span> <span class="mi">1</span><span class="p">))</span>
- <span class="n">losses</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="p">(</span><span class="nb">len</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">valid_dl</span><span class="p">),</span> <span class="mi">1</span><span class="p">))</span>
-
- <span class="c1"># iterate over the validation/test set</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Calculating accuracy on the validation set ...'</span><span class="p">)</span>
- <span class="k">for</span> <span class="n">batch</span><span class="p">,</span> <span class="p">(</span><span class="n">inputs</span><span class="p">,</span> <span class="n">labels</span><span class="p">)</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">valid_dl</span><span class="p">):</span>
-
- <span class="c1"># send the data to the gpu if available</span>
- <span class="n">inputs</span> <span class="o">=</span> <span class="n">inputs</span><span class="o">.</span><span class="n">to</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">device</span><span class="p">)</span>
- <span class="n">labels</span> <span class="o">=</span> <span class="n">labels</span><span class="o">.</span><span class="n">to</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">device</span><span class="p">)</span>
-
- <span class="c1"># calculate network outputs</span>
- <span class="k">with</span> <span class="n">torch</span><span class="o">.</span><span class="n">no_grad</span><span class="p">():</span>
- <span class="n">outputs</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">model</span><span class="p">(</span><span class="n">inputs</span><span class="p">)</span>
-
- <span class="c1"># compute loss</span>
- <span class="n">loss</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">loss_function</span><span class="p">(</span><span class="n">outputs</span><span class="p">,</span> <span class="n">labels</span><span class="o">.</span><span class="n">long</span><span class="p">())</span>
- <span class="n">losses</span><span class="p">[</span><span class="n">batch</span><span class="p">,</span> <span class="mi">0</span><span class="p">]</span> <span class="o">=</span> <span class="n">loss</span><span class="o">.</span><span class="n">detach</span><span class="p">()</span><span class="o">.</span><span class="n">numpy</span><span class="p">()</span><span class="o">.</span><span class="n">item</span><span class="p">()</span>
-
- <span class="c1"># calculate predicted class labels</span>
- <span class="n">pred</span> <span class="o">=</span> <span class="n">F</span><span class="o">.</span><span class="n">softmax</span><span class="p">(</span><span class="n">outputs</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="mi">1</span><span class="p">)</span><span class="o">.</span><span class="n">argmax</span><span class="p">(</span><span class="n">dim</span><span class="o">=</span><span class="mi">1</span><span class="p">)</span>
-
- <span class="c1"># calculate accuracy on current batch</span>
- <span class="n">acc</span> <span class="o">=</span> <span class="n">accuracy_function</span><span class="p">(</span><span class="n">pred</span><span class="p">,</span> <span class="n">labels</span><span class="p">)</span>
- <span class="n">accuracies</span><span class="p">[</span><span class="n">batch</span><span class="p">,</span> <span class="mi">0</span><span class="p">]</span> <span class="o">=</span> <span class="n">acc</span>
-
- <span class="c1"># print progress</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Mini-batch: </span><span class="si">{:d}</span><span class="s1">/</span><span class="si">{:d}</span><span class="s1">, Accuracy: </span><span class="si">{:.2f}</span><span class="s1">'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">batch</span> <span class="o">+</span> <span class="mi">1</span><span class="p">,</span> <span class="nb">len</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">valid_dl</span><span class="p">),</span> <span class="n">acc</span><span class="p">))</span>
-
- <span class="c1"># calculate overall accuracy on the validation/test set</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Epoch: </span><span class="si">{:d}</span><span class="s1">, Mean accuracy: </span><span class="si">{:.2f}</span><span class="s1">%.'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">model</span><span class="o">.</span><span class="n">epoch</span><span class="p">,</span> <span class="n">accuracies</span><span class="o">.</span><span class="n">mean</span><span class="p">()</span> <span class="o">*</span> <span class="mi">100</span><span class="p">))</span>
-
- <span class="k">return</span> <span class="n">accuracies</span><span class="p">,</span> <span class="n">losses</span></div>
-
-<div class="viewcode-block" id="NetworkTrainer.save_state"><a class="viewcode-back" href="../../source/core.html#core.trainer.NetworkTrainer.save_state">[docs]</a> <span class="k">def</span> <span class="nf">save_state</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Save the model state.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># whether to save the model state</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">save</span><span class="p">:</span>
-
- <span class="c1"># append the model performance before the checkpoint to the model</span>
- <span class="c1"># state, if a checkpoint is passed</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">checkpoint_state</span><span class="p">:</span>
-
- <span class="c1"># append values from checkpoint to current training state</span>
- <span class="n">state</span> <span class="o">=</span> <span class="p">{</span><span class="n">k1</span><span class="p">:</span> <span class="n">np</span><span class="o">.</span><span class="n">hstack</span><span class="p">([</span><span class="n">v1</span><span class="p">,</span> <span class="n">v2</span><span class="p">])</span> <span class="k">for</span> <span class="p">(</span><span class="n">k1</span><span class="p">,</span> <span class="n">v1</span><span class="p">),</span> <span class="p">(</span><span class="n">k2</span><span class="p">,</span> <span class="n">v2</span><span class="p">)</span> <span class="ow">in</span>
- <span class="nb">zip</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">checkpoint_state</span><span class="o">.</span><span class="n">items</span><span class="p">(),</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">training_state</span><span class="o">.</span><span class="n">items</span><span class="p">())</span> <span class="k">if</span> <span class="n">k1</span> <span class="o">==</span> <span class="n">k2</span><span class="p">}</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="n">state</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">training_state</span>
-
- <span class="c1"># save model state</span>
- <span class="n">_</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">model</span><span class="o">.</span><span class="n">save</span><span class="p">(</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">state_file</span><span class="p">,</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">optimizer</span><span class="p">,</span>
- <span class="n">bands</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">train_dl</span><span class="o">.</span><span class="n">dataset</span><span class="o">.</span><span class="n">dataset</span><span class="o">.</span><span class="n">use_bands</span><span class="p">,</span>
- <span class="n">train_ds</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">train_dl</span><span class="o">.</span><span class="n">dataset</span><span class="p">,</span>
- <span class="n">valid_ds</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">valid_dl</span><span class="o">.</span><span class="n">dataset</span><span class="p">,</span>
- <span class="n">test_ds</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">test_dl</span><span class="o">.</span><span class="n">dataset</span><span class="p">,</span>
- <span class="n">state</span><span class="o">=</span><span class="n">state</span><span class="p">,</span>
- <span class="p">)</span></div>
-
- <span class="k">def</span> <span class="fm">__repr__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Representation of `~pysegcnn.core.trainer.NetworkTrainer`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> fs : `str`</span>
-<span class="sd"> Representation string.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># representation string to print</span>
- <span class="n">fs</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="vm">__class__</span><span class="o">.</span><span class="vm">__name__</span> <span class="o">+</span> <span class="s1">'(</span><span class="se">\n</span><span class="s1">'</span>
-
- <span class="c1"># dataset</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">' (dataset):</span><span class="se">\n</span><span class="s1"> '</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">''</span><span class="o">.</span><span class="n">join</span><span class="p">(</span>
- <span class="nb">repr</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">train_dl</span><span class="o">.</span><span class="n">dataset</span><span class="o">.</span><span class="n">dataset</span><span class="p">))</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="s1">'</span><span class="se">\n</span><span class="s1">'</span><span class="p">,</span> <span class="s1">'</span><span class="se">\n</span><span class="s1"> '</span><span class="p">)</span>
-
- <span class="c1"># batch size</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">'</span><span class="se">\n</span><span class="s1"> (batch):</span><span class="se">\n</span><span class="s1"> '</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">'- batch size: </span><span class="si">{}</span><span class="se">\n</span><span class="s1"> '</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">train_dl</span><span class="o">.</span><span class="n">batch_size</span><span class="p">)</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">'- mini-batch shape (b, c, h, w): </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span>
- <span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">train_dl</span><span class="o">.</span><span class="n">batch_size</span><span class="p">,</span>
- <span class="nb">len</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">train_dl</span><span class="o">.</span><span class="n">dataset</span><span class="o">.</span><span class="n">dataset</span><span class="o">.</span><span class="n">use_bands</span><span class="p">),</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">train_dl</span><span class="o">.</span><span class="n">dataset</span><span class="o">.</span><span class="n">dataset</span><span class="o">.</span><span class="n">tile_size</span><span class="p">,</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">train_dl</span><span class="o">.</span><span class="n">dataset</span><span class="o">.</span><span class="n">dataset</span><span class="o">.</span><span class="n">tile_size</span><span class="p">)</span>
- <span class="p">)</span>
-
- <span class="c1"># dataset split</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">'</span><span class="se">\n</span><span class="s1"> (split):'</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">'</span><span class="se">\n</span><span class="s1"> '</span> <span class="o">+</span> <span class="nb">repr</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">train_dl</span><span class="o">.</span><span class="n">dataset</span><span class="p">)</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">'</span><span class="se">\n</span><span class="s1"> '</span> <span class="o">+</span> <span class="nb">repr</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">valid_dl</span><span class="o">.</span><span class="n">dataset</span><span class="p">)</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">'</span><span class="se">\n</span><span class="s1"> '</span> <span class="o">+</span> <span class="nb">repr</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">test_dl</span><span class="o">.</span><span class="n">dataset</span><span class="p">)</span>
-
- <span class="c1"># model</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">'</span><span class="se">\n</span><span class="s1"> (model):</span><span class="se">\n</span><span class="s1"> '</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">''</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="nb">repr</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">model</span><span class="p">))</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="s1">'</span><span class="se">\n</span><span class="s1">'</span><span class="p">,</span> <span class="s1">'</span><span class="se">\n</span><span class="s1"> '</span><span class="p">)</span>
-
- <span class="c1"># optimizer</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">'</span><span class="se">\n</span><span class="s1"> (optimizer):</span><span class="se">\n</span><span class="s1"> '</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">''</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="nb">repr</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">optimizer</span><span class="p">))</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="s1">'</span><span class="se">\n</span><span class="s1">'</span><span class="p">,</span> <span class="s1">'</span><span class="se">\n</span><span class="s1"> '</span><span class="p">)</span>
-
- <span class="c1"># early stopping</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">'</span><span class="se">\n</span><span class="s1"> (early stop):</span><span class="se">\n</span><span class="s1"> '</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">''</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="nb">repr</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">es</span><span class="p">))</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="s1">'</span><span class="se">\n</span><span class="s1">'</span><span class="p">,</span> <span class="s1">'</span><span class="se">\n</span><span class="s1"> '</span><span class="p">)</span>
-
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">'</span><span class="se">\n</span><span class="s1">)'</span>
- <span class="k">return</span> <span class="n">fs</span></div>
-
-
-<div class="viewcode-block" id="EarlyStopping"><a class="viewcode-back" href="../../source/core.html#core.trainer.EarlyStopping">[docs]</a><span class="k">class</span> <span class="nc">EarlyStopping</span><span class="p">(</span><span class="nb">object</span><span class="p">):</span>
- <span class="sd">"""`Early stopping`_ algorithm.</span>
-
-<span class="sd"> This implementation of the early stopping algorithm advances a counter each</span>
-<span class="sd"> time a metric did not improve over a training epoch. If the metric does not</span>
-<span class="sd"> improve over more than ``patience`` epochs, the early stopping criterion is</span>
-<span class="sd"> met.</span>
-
-<span class="sd"> See `pysegcnn.core.trainer.NetworkTrainer.train` for an example</span>
-<span class="sd"> implementation.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> mode : `str`, optional</span>
-<span class="sd"> The mode of the early stopping. Depends on the metric measuring</span>
-<span class="sd"> performance. When using model loss as metric, use ``mode`` = 'min',</span>
-<span class="sd"> however, when using accuracy as metric, use ``mode`` = 'max'. The</span>
-<span class="sd"> default is 'max'.</span>
-<span class="sd"> best : `float`, optional</span>
-<span class="sd"> Threshold indicating the best metric score. At instanciation, set</span>
-<span class="sd"> ``best`` to the worst possible score of the metric. ``best`` will be</span>
-<span class="sd"> overwritten during training. The default is 0.</span>
-<span class="sd"> min_delta : `float`, optional</span>
-<span class="sd"> Minimum change in early stopping metric to be considered as an</span>
-<span class="sd"> improvement. The default is 0.</span>
-<span class="sd"> patience : `int`, optional</span>
-<span class="sd"> The number of epochs to wait for an improvement in the early stopping</span>
-<span class="sd"> metric. The default is 10.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> ValueError</span>
-<span class="sd"> Raised if ``mode`` is not either 'min' or 'max'.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> .. _Early stopping:</span>
-<span class="sd"> https://en.wikipedia.org/wiki/Early_stopping</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">mode</span><span class="o">=</span><span class="s1">'max'</span><span class="p">,</span> <span class="n">best</span><span class="o">=</span><span class="mi">0</span><span class="p">,</span> <span class="n">min_delta</span><span class="o">=</span><span class="mi">0</span><span class="p">,</span> <span class="n">patience</span><span class="o">=</span><span class="mi">10</span><span class="p">):</span>
-
- <span class="c1"># check if mode is correctly specified</span>
- <span class="k">if</span> <span class="n">mode</span> <span class="ow">not</span> <span class="ow">in</span> <span class="p">[</span><span class="s1">'min'</span><span class="p">,</span> <span class="s1">'max'</span><span class="p">]:</span>
- <span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="s1">'Mode "</span><span class="si">{}</span><span class="s1">" not supported. '</span>
- <span class="s1">'Mode is either "min" (check whether the metric '</span>
- <span class="s1">'decreased, e.g. loss) or "max" (check whether '</span>
- <span class="s1">'the metric increased, e.g. accuracy).'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">mode</span><span class="p">))</span>
-
- <span class="c1"># mode to determine if metric improved</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">mode</span> <span class="o">=</span> <span class="n">mode</span>
-
- <span class="c1"># whether to check for an increase or a decrease in a given metric</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">is_better</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">decreased</span> <span class="k">if</span> <span class="n">mode</span> <span class="o">==</span> <span class="s1">'min'</span> <span class="k">else</span> <span class="bp">self</span><span class="o">.</span><span class="n">increased</span>
-
- <span class="c1"># minimum change in metric to be considered as an improvement</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">min_delta</span> <span class="o">=</span> <span class="n">min_delta</span>
-
- <span class="c1"># number of epochs to wait for improvement</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">patience</span> <span class="o">=</span> <span class="n">patience</span>
-
- <span class="c1"># initialize best metric</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">best</span> <span class="o">=</span> <span class="n">best</span>
-
- <span class="c1"># initialize early stopping flag</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">early_stop</span> <span class="o">=</span> <span class="kc">False</span>
-
- <span class="c1"># initialize the early stop counter</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">counter</span> <span class="o">=</span> <span class="mi">0</span>
-
-<div class="viewcode-block" id="EarlyStopping.stop"><a class="viewcode-back" href="../../source/core.html#core.trainer.EarlyStopping.stop">[docs]</a> <span class="k">def</span> <span class="nf">stop</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">metric</span><span class="p">):</span>
- <span class="sd">"""Advance early stopping counter.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> metric : `float`</span>
-<span class="sd"> The current metric score.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> early_stop : `bool`</span>
-<span class="sd"> Whether the early stopping criterion is met.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># if the metric improved, reset the epochs counter, else, advance</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">is_better</span><span class="p">(</span><span class="n">metric</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">best</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">min_delta</span><span class="p">):</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">counter</span> <span class="o">=</span> <span class="mi">0</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">best</span> <span class="o">=</span> <span class="n">metric</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">counter</span> <span class="o">+=</span> <span class="mi">1</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Early stopping counter: </span><span class="si">{}</span><span class="s1">/</span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">counter</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">patience</span><span class="p">))</span>
-
- <span class="c1"># if the metric did not improve over the last patience epochs,</span>
- <span class="c1"># the early stopping criterion is met</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">counter</span> <span class="o">>=</span> <span class="bp">self</span><span class="o">.</span><span class="n">patience</span><span class="p">:</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Early stopping criterion met, stopping training.'</span><span class="p">)</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">early_stop</span> <span class="o">=</span> <span class="kc">True</span>
-
- <span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">early_stop</span></div>
-
-<div class="viewcode-block" id="EarlyStopping.decreased"><a class="viewcode-back" href="../../source/core.html#core.trainer.EarlyStopping.decreased">[docs]</a> <span class="k">def</span> <span class="nf">decreased</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">metric</span><span class="p">,</span> <span class="n">best</span><span class="p">,</span> <span class="n">min_delta</span><span class="p">):</span>
- <span class="sd">"""Whether a metric decreased with respect to a best score.</span>
-
-<span class="sd"> Measure improvement for metrics that are considered as 'better' when</span>
-<span class="sd"> they decrease, e.g. model loss, mean squared error, etc.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> metric : `float`</span>
-<span class="sd"> The current score.</span>
-<span class="sd"> best : `float`</span>
-<span class="sd"> The current best score.</span>
-<span class="sd"> min_delta : `float`</span>
-<span class="sd"> Minimum change to be considered as an improvement.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> `bool`</span>
-<span class="sd"> Whether the metric improved.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="n">metric</span> <span class="o"><</span> <span class="n">best</span> <span class="o">-</span> <span class="n">min_delta</span></div>
-
-<div class="viewcode-block" id="EarlyStopping.increased"><a class="viewcode-back" href="../../source/core.html#core.trainer.EarlyStopping.increased">[docs]</a> <span class="k">def</span> <span class="nf">increased</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">metric</span><span class="p">,</span> <span class="n">best</span><span class="p">,</span> <span class="n">min_delta</span><span class="p">):</span>
- <span class="sd">"""Whether a metric increased with respect to a best score.</span>
-
-<span class="sd"> Measure improvement for metrics that are considered as 'better' when</span>
-<span class="sd"> they increase, e.g. accuracy, precision, recall, etc.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> metric : `float`</span>
-<span class="sd"> The current score.</span>
-<span class="sd"> best : `float`</span>
-<span class="sd"> The current best score.</span>
-<span class="sd"> min_delta : `float`</span>
-<span class="sd"> Minimum change to be considered as an improvement.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> `bool`</span>
-<span class="sd"> Whether the metric improved.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="n">metric</span> <span class="o">></span> <span class="n">best</span> <span class="o">+</span> <span class="n">min_delta</span></div>
-
- <span class="k">def</span> <span class="fm">__repr__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Representation of `~pysegcnn.core.trainer.EarlyStopping`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> fs : `str`</span>
-<span class="sd"> Representation string.</span>
-
-<span class="sd"> """</span>
- <span class="n">fs</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="vm">__class__</span><span class="o">.</span><span class="vm">__name__</span>
- <span class="n">fs</span> <span class="o">+=</span> <span class="s1">'(mode=</span><span class="si">{}</span><span class="s1">, best=</span><span class="si">{:.2f}</span><span class="s1">, delta=</span><span class="si">{}</span><span class="s1">, patience=</span><span class="si">{}</span><span class="s1">)'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">mode</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">best</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">min_delta</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">patience</span><span class="p">)</span>
-
- <span class="k">return</span> <span class="n">fs</span></div>
-</pre></div>
-
- </div>
-
- </div>
- <footer>
-
-
- <hr/>
-
- <div role="contentinfo">
- <p>
-
- © Copyright 2020, Daniel Frisinghelli
-
- </p>
- </div>
-
-
-
- Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a
-
- <a href="https://github.com/rtfd/sphinx_rtd_theme">theme</a>
-
- provided by <a href="https://readthedocs.org">Read the Docs</a>.
-
-</footer>
-
- </div>
- </div>
-
- </section>
-
- </div>
-
-
- <script type="text/javascript">
- jQuery(function () {
- SphinxRtdTheme.Navigation.enable(true);
- });
- </script>
-
-
-
-
-
-
-</body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/_modules/core/transforms.html b/docs/_build/html/_modules/core/transforms.html
deleted file mode 100644
index 51ff97a3d4915ed13bf050559143b3c215e6d801..0000000000000000000000000000000000000000
--- a/docs/_build/html/_modules/core/transforms.html
+++ /dev/null
@@ -1,647 +0,0 @@
-
-
-<!DOCTYPE html>
-<html class="writer-html5" lang="en" >
-<head>
- <meta charset="utf-8">
-
- <meta name="viewport" content="width=device-width, initial-scale=1.0">
-
- <title>core.transforms — PySegCNN documentation</title>
-
-
-
- <link rel="stylesheet" href="../../_static/css/theme.css" type="text/css" />
- <link rel="stylesheet" href="../../_static/pygments.css" type="text/css" />
-
-
-
-
-
-
-
- <!--[if lt IE 9]>
- <script src="../../_static/js/html5shiv.min.js"></script>
- <![endif]-->
-
-
- <script type="text/javascript" id="documentation_options" data-url_root="../../" src="../../_static/documentation_options.js"></script>
- <script src="../../_static/jquery.js"></script>
- <script src="../../_static/underscore.js"></script>
- <script src="../../_static/doctools.js"></script>
- <script src="../../_static/language_data.js"></script>
-
- <script type="text/javascript" src="../../_static/js/theme.js"></script>
-
-
- <link rel="index" title="Index" href="../../genindex.html" />
- <link rel="search" title="Search" href="../../search.html" />
-</head>
-
-<body class="wy-body-for-nav">
-
-
- <div class="wy-grid-for-nav">
-
- <nav data-toggle="wy-nav-shift" class="wy-nav-side">
- <div class="wy-side-scroll">
- <div class="wy-side-nav-search" >
-
-
-
- <a href="../../index.html" class="icon icon-home" alt="Documentation Home"> PySegCNN
-
-
-
- </a>
-
-
-
-
-
-
-
-<div role="search">
- <form id="rtd-search-form" class="wy-form" action="../../search.html" method="get">
- <input type="text" name="q" placeholder="Search docs" />
- <input type="hidden" name="check_keywords" value="yes" />
- <input type="hidden" name="area" value="default" />
- </form>
-</div>
-
-
- </div>
-
-
- <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
-
-
-
-
-
-
- <ul>
-<li class="toctree-l1"><a class="reference internal" href="../../source/requirements.html">Requirements</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../../source/installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../../source/api.html">API Reference</a></li>
-</ul>
-
-
-
- </div>
-
- </div>
- </nav>
-
- <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
-
-
- <nav class="wy-nav-top" aria-label="top navigation">
-
- <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
- <a href="../../index.html">PySegCNN</a>
-
- </nav>
-
-
- <div class="wy-nav-content">
-
- <div class="rst-content">
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-<div role="navigation" aria-label="breadcrumbs navigation">
-
- <ul class="wy-breadcrumbs">
-
- <li><a href="../../index.html" class="icon icon-home"></a> »</li>
-
- <li><a href="../index.html">Module code</a> »</li>
-
- <li>core.transforms</li>
-
-
- <li class="wy-breadcrumbs-aside">
-
- </li>
-
- </ul>
-
-
- <hr/>
-</div>
- <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
- <div itemprop="articleBody">
-
- <h1>Source code for core.transforms</h1><div class="highlight"><pre>
-<span></span><span class="sd">"""Data augmentation.</span>
-
-<span class="sd">This module provides classes implementing common image augmentation methods.</span>
-
-<span class="sd">These methods may be used to artificially increase a dataset.</span>
-
-<span class="sd">License</span>
-<span class="sd">-------</span>
-
-<span class="sd"> Copyright (c) 2020 Daniel Frisinghelli</span>
-
-<span class="sd"> This source code is licensed under the GNU General Public License v3.</span>
-
-<span class="sd"> See the LICENSE file in the repository's root directory.</span>
-
-<span class="sd">"""</span>
-
-<span class="c1"># !/usr/bin/env python</span>
-<span class="c1"># -*- coding: utf-8 -*-</span>
-
-<span class="c1"># externals</span>
-<span class="kn">import</span> <span class="nn">numpy</span> <span class="k">as</span> <span class="nn">np</span>
-<span class="kn">from</span> <span class="nn">scipy</span> <span class="kn">import</span> <span class="n">ndimage</span>
-
-
-<div class="viewcode-block" id="Transform"><a class="viewcode-back" href="../../source/core.html#core.transforms.Transform">[docs]</a><span class="k">class</span> <span class="nc">Transform</span><span class="p">(</span><span class="nb">object</span><span class="p">):</span>
- <span class="sd">"""Base class for an image transformation."""</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="k">pass</span>
-
- <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">image</span><span class="p">):</span>
- <span class="sd">"""Apply transformation.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> image : `numpy.ndarray`</span>
-<span class="sd"> The image to transform.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> NotImplementedError</span>
-<span class="sd"> Raised if `~pysegcnn.core.transforms.Transform` is not inherited.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
- <span class="k">raise</span> <span class="ne">NotImplementedError</span></div>
-
-
-<div class="viewcode-block" id="VariantTransform"><a class="viewcode-back" href="../../source/core.html#core.transforms.VariantTransform">[docs]</a><span class="k">class</span> <span class="nc">VariantTransform</span><span class="p">(</span><span class="n">Transform</span><span class="p">):</span>
- <span class="sd">"""Base class for a spatially variant transformation.</span>
-
-<span class="sd"> Transformation on the ground truth required.</span>
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
-
- <span class="c1"># transformation on the ground truth required</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">invariant</span> <span class="o">=</span> <span class="kc">False</span></div>
-
-
-<div class="viewcode-block" id="InvariantTransform"><a class="viewcode-back" href="../../source/core.html#core.transforms.InvariantTransform">[docs]</a><span class="k">class</span> <span class="nc">InvariantTransform</span><span class="p">(</span><span class="n">Transform</span><span class="p">):</span>
- <span class="sd">"""Base class for a spatially invariant transformation.</span>
-
-<span class="sd"> Transformation on the ground truth not required.</span>
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
-
- <span class="c1"># transformation on the ground truth not required</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">invariant</span> <span class="o">=</span> <span class="kc">True</span></div>
-
-
-<div class="viewcode-block" id="FlipLr"><a class="viewcode-back" href="../../source/core.html#core.transforms.FlipLr">[docs]</a><span class="k">class</span> <span class="nc">FlipLr</span><span class="p">(</span><span class="n">VariantTransform</span><span class="p">):</span>
- <span class="sd">"""Flip an image horizontally.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> p : `float`, optional</span>
-<span class="sd"> The probability to apply the transformation. The default is 0.5.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">p</span><span class="o">=</span><span class="mf">0.5</span><span class="p">):</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">()</span>
- <span class="c1"># the probability to apply the transformation</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">p</span> <span class="o">=</span> <span class="n">p</span>
-
- <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">image</span><span class="p">):</span>
- <span class="sd">"""Apply transformation.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> image : `numpy.ndarray`</span>
-<span class="sd"> The image to transform.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> transform : `numpy.ndarray`</span>
-<span class="sd"> The transformed image.</span>
-
-<span class="sd"> """</span>
- <span class="k">if</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">random</span><span class="p">(</span><span class="mi">1</span><span class="p">)</span> <span class="o"><</span> <span class="bp">self</span><span class="o">.</span><span class="n">p</span><span class="p">:</span>
- <span class="c1"># transformation applied</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">applied</span> <span class="o">=</span> <span class="kc">True</span>
- <span class="k">return</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">image</span><span class="p">)[</span><span class="o">...</span><span class="p">,</span> <span class="p">::</span><span class="o">-</span><span class="mi">1</span><span class="p">]</span>
-
- <span class="c1"># transformation not applied</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">applied</span> <span class="o">=</span> <span class="kc">False</span>
- <span class="k">return</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">image</span><span class="p">)</span>
-
- <span class="k">def</span> <span class="fm">__repr__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Representation of the transformation.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> repr : `str`</span>
-<span class="sd"> Representation string.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="vm">__class__</span><span class="o">.</span><span class="vm">__name__</span> <span class="o">+</span> <span class="s1">'(p = </span><span class="si">{}</span><span class="s1">)'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">p</span><span class="p">)</span></div>
-
-
-<div class="viewcode-block" id="FlipUd"><a class="viewcode-back" href="../../source/core.html#core.transforms.FlipUd">[docs]</a><span class="k">class</span> <span class="nc">FlipUd</span><span class="p">(</span><span class="n">VariantTransform</span><span class="p">):</span>
- <span class="sd">"""Flip an image vertically.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> p : `float`, optional</span>
-<span class="sd"> The probability to apply the transformation. The default is 0.5.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">p</span><span class="o">=</span><span class="mf">0.5</span><span class="p">):</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">()</span>
- <span class="c1"># the probability to apply the transformation</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">p</span> <span class="o">=</span> <span class="n">p</span>
-
- <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">image</span><span class="p">):</span>
- <span class="sd">"""Apply transformation.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> image : `numpy.ndarray`</span>
-<span class="sd"> The image to transform.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> transform : `numpy.ndarray`</span>
-<span class="sd"> The transformed image.</span>
-
-<span class="sd"> """</span>
- <span class="k">if</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">random</span><span class="p">(</span><span class="mi">1</span><span class="p">)</span> <span class="o"><</span> <span class="bp">self</span><span class="o">.</span><span class="n">p</span><span class="p">:</span>
- <span class="c1"># transformation applied</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">applied</span> <span class="o">=</span> <span class="kc">True</span>
- <span class="k">return</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">image</span><span class="p">)[</span><span class="o">...</span><span class="p">,</span> <span class="p">::</span><span class="o">-</span><span class="mi">1</span><span class="p">,</span> <span class="p">:]</span>
-
- <span class="c1"># transformation not applied</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">applied</span> <span class="o">=</span> <span class="kc">False</span>
- <span class="k">return</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">image</span><span class="p">)</span>
-
- <span class="k">def</span> <span class="fm">__repr__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Representation of the transformation.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> repr : `str`</span>
-<span class="sd"> Representation string.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="vm">__class__</span><span class="o">.</span><span class="vm">__name__</span> <span class="o">+</span> <span class="s1">'(p = </span><span class="si">{}</span><span class="s1">)'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">p</span><span class="p">)</span></div>
-
-
-<div class="viewcode-block" id="Rotate"><a class="viewcode-back" href="../../source/core.html#core.transforms.Rotate">[docs]</a><span class="k">class</span> <span class="nc">Rotate</span><span class="p">(</span><span class="n">VariantTransform</span><span class="p">):</span>
- <span class="sd">"""Rotate an image by ``angle``.</span>
-
-<span class="sd"> The image is rotated in the spatial plane.</span>
-
-<span class="sd"> If the input array has more then two dimensions, the spatial dimensions are</span>
-<span class="sd"> assumed to be the last two dimensions of the array.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> angle : `float`</span>
-<span class="sd"> The rotation angle in degrees.</span>
-<span class="sd"> p : `float`, optional</span>
-<span class="sd"> The probability to apply the transformation. The default is 0.5.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">angle</span><span class="p">,</span> <span class="n">p</span><span class="o">=</span><span class="mf">0.5</span><span class="p">):</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">()</span>
-
- <span class="c1"># the rotation angle</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">angle</span> <span class="o">=</span> <span class="n">angle</span>
-
- <span class="c1"># the probability to apply the transformation</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">p</span> <span class="o">=</span> <span class="n">p</span>
-
- <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">image</span><span class="p">):</span>
- <span class="sd">"""Apply transformation.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> image : `numpy.ndarray`</span>
-<span class="sd"> The image to transform.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> transform : `numpy.ndarray`</span>
-<span class="sd"> The transformed image.</span>
-<span class="sd"> """</span>
- <span class="k">if</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">random</span><span class="p">(</span><span class="mi">1</span><span class="p">)</span> <span class="o"><</span> <span class="bp">self</span><span class="o">.</span><span class="n">p</span><span class="p">:</span>
-
- <span class="c1"># transformation applied</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">applied</span> <span class="o">=</span> <span class="kc">True</span>
-
- <span class="c1"># check dimension of input image</span>
- <span class="n">ndim</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">image</span><span class="p">)</span><span class="o">.</span><span class="n">ndim</span>
-
- <span class="c1"># axes defining the rotational plane</span>
- <span class="n">rot_axes</span> <span class="o">=</span> <span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="mi">1</span><span class="p">)</span>
- <span class="k">if</span> <span class="n">ndim</span> <span class="o">></span> <span class="mi">2</span><span class="p">:</span>
- <span class="n">rot_axes</span> <span class="o">=</span> <span class="p">(</span><span class="n">ndim</span> <span class="o">-</span> <span class="mi">2</span><span class="p">,</span> <span class="n">ndim</span> <span class="o">-</span> <span class="mi">1</span><span class="p">)</span>
-
- <span class="k">return</span> <span class="n">ndimage</span><span class="o">.</span><span class="n">rotate</span><span class="p">(</span><span class="n">image</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">angle</span><span class="p">,</span> <span class="n">axes</span><span class="o">=</span><span class="n">rot_axes</span><span class="p">,</span>
- <span class="n">reshape</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
-
- <span class="c1"># transformation not applied</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">applied</span> <span class="o">=</span> <span class="kc">False</span>
- <span class="k">return</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">image</span><span class="p">)</span>
-
- <span class="k">def</span> <span class="fm">__repr__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Representation of the transformation.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> repr : `str`</span>
-<span class="sd"> Representation string.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="vm">__class__</span><span class="o">.</span><span class="vm">__name__</span> <span class="o">+</span> <span class="s1">'(angle = </span><span class="si">{}</span><span class="s1">, p = </span><span class="si">{}</span><span class="s1">)'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">angle</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">p</span><span class="p">)</span></div>
-
-
-<div class="viewcode-block" id="Noise"><a class="viewcode-back" href="../../source/core.html#core.transforms.Noise">[docs]</a><span class="k">class</span> <span class="nc">Noise</span><span class="p">(</span><span class="n">InvariantTransform</span><span class="p">):</span>
- <span class="sd">"""Add gaussian noise to an image.</span>
-
-<span class="sd"> Valid modes are:</span>
-
-<span class="sd"> 'add': image = image + noise</span>
-<span class="sd"> 'speckle' : image = image + image * noise</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> mode : `str`</span>
-<span class="sd"> The mode to add the noise.</span>
-<span class="sd"> mean : `float`, optional</span>
-<span class="sd"> The mean of the gaussian distribution from which the noise is sampled.</span>
-<span class="sd"> The default is 0.</span>
-<span class="sd"> var : `float`, optional</span>
-<span class="sd"> The variance of the gaussian distribution from which the noise is</span>
-<span class="sd"> sampled. The default is 0.05.</span>
-<span class="sd"> p : `float`, optional</span>
-<span class="sd"> The probability to apply the transformation. The default is 0.5.</span>
-<span class="sd"> exclude : `list` [`float`] or `list` [`int`], optional</span>
-<span class="sd"> Values for which the noise is not added. Useful for pixels resulting</span>
-<span class="sd"> from image padding. The default is [].</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> ValueError</span>
-<span class="sd"> Raised if ``mode`` is not supported.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="c1"># supported modes</span>
- <span class="n">modes</span> <span class="o">=</span> <span class="p">[</span><span class="s1">'add'</span><span class="p">,</span> <span class="s1">'speckle'</span><span class="p">]</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">mode</span><span class="p">,</span> <span class="n">mean</span><span class="o">=</span><span class="mi">0</span><span class="p">,</span> <span class="n">var</span><span class="o">=</span><span class="mf">0.05</span><span class="p">,</span> <span class="n">p</span><span class="o">=</span><span class="mf">0.5</span><span class="p">,</span> <span class="n">exclude</span><span class="o">=</span><span class="p">[]):</span>
- <span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">()</span>
-
- <span class="c1"># check which kind of noise to apply</span>
- <span class="k">if</span> <span class="n">mode</span> <span class="ow">not</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">modes</span><span class="p">:</span>
- <span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="s1">'Supported noise types are: </span><span class="si">{}</span><span class="s1">.'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">modes</span><span class="p">))</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">mode</span> <span class="o">=</span> <span class="n">mode</span>
-
- <span class="c1"># mean and variance of the gaussian distribution the noise signal is</span>
- <span class="c1"># sampled from</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">mean</span> <span class="o">=</span> <span class="n">mean</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">var</span> <span class="o">=</span> <span class="n">var</span>
-
- <span class="c1"># the probability to apply the transformation</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">p</span> <span class="o">=</span> <span class="n">p</span>
-
- <span class="c1"># the value in the image that should not be modified by the noise</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">exclude</span> <span class="o">=</span> <span class="n">exclude</span>
-
- <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">image</span><span class="p">):</span>
- <span class="sd">"""Apply transformation.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> image : `numpy.ndarray`</span>
-<span class="sd"> The image to transform.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> transform : `numpy.ndarray`</span>
-<span class="sd"> The transformed image</span>
-
-<span class="sd"> """</span>
- <span class="k">if</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">random</span><span class="p">(</span><span class="mi">1</span><span class="p">)</span> <span class="o"><</span> <span class="bp">self</span><span class="o">.</span><span class="n">p</span><span class="p">:</span>
-
- <span class="c1"># transformation applied</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">applied</span> <span class="o">=</span> <span class="kc">True</span>
-
- <span class="c1"># generate gaussian noise</span>
- <span class="n">noise</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">normal</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">mean</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">var</span><span class="p">,</span> <span class="n">image</span><span class="o">.</span><span class="n">shape</span><span class="p">)</span>
-
- <span class="c1"># check which values should not be modified by adding noise</span>
- <span class="k">for</span> <span class="n">val</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">exclude</span><span class="p">:</span>
- <span class="n">noise</span><span class="p">[</span><span class="n">image</span> <span class="o">==</span> <span class="n">val</span><span class="p">]</span> <span class="o">=</span> <span class="mi">0</span>
-
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">mode</span> <span class="o">==</span> <span class="s1">'gaussian'</span><span class="p">:</span>
- <span class="k">return</span> <span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">image</span><span class="p">)</span> <span class="o">+</span> <span class="n">noise</span><span class="p">)</span><span class="o">.</span><span class="n">clip</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="mi">1</span><span class="p">)</span>
-
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">mode</span> <span class="o">==</span> <span class="s1">'speckle'</span><span class="p">:</span>
- <span class="k">return</span> <span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">image</span><span class="p">)</span> <span class="o">+</span>
- <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">image</span><span class="p">)</span> <span class="o">*</span> <span class="n">noise</span><span class="p">)</span><span class="o">.</span><span class="n">clip</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="mi">1</span><span class="p">)</span>
-
- <span class="c1"># transformation not applied</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">applied</span> <span class="o">=</span> <span class="kc">False</span>
- <span class="k">return</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">image</span><span class="p">)</span>
-
- <span class="k">def</span> <span class="fm">__repr__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Representation of the transformation.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> repr : `str`</span>
-<span class="sd"> Representation string.</span>
-
-<span class="sd"> """</span>
- <span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="vm">__class__</span><span class="o">.</span><span class="vm">__name__</span> <span class="o">+</span> <span class="p">(</span><span class="s1">'(mode = </span><span class="si">{}</span><span class="s1">, mean = </span><span class="si">{}</span><span class="s1">, var = </span><span class="si">{}</span><span class="s1">, '</span>
- <span class="s1">'p = </span><span class="si">{}</span><span class="s1">)'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">mode</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">mean</span><span class="p">,</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">var</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">p</span><span class="p">))</span></div>
-
-
-<div class="viewcode-block" id="Augment"><a class="viewcode-back" href="../../source/core.html#core.transforms.Augment">[docs]</a><span class="k">class</span> <span class="nc">Augment</span><span class="p">(</span><span class="nb">object</span><span class="p">):</span>
- <span class="sd">"""Apply a sequence of transformations.</span>
-
-<span class="sd"> Container class applying each transformation in ``transforms`` in order.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> transforms : `list` or `tuple`</span>
-<span class="sd"> A sequence of instances of `pysegcnn.core.transforms.VariantTransform`</span>
-<span class="sd"> or `pysegcnn.core.transforms.InvariantTransform`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
-
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">transforms</span><span class="p">):</span>
- <span class="k">assert</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">transforms</span><span class="p">,</span> <span class="p">(</span><span class="nb">list</span><span class="p">,</span> <span class="nb">tuple</span><span class="p">))</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">transforms</span> <span class="o">=</span> <span class="n">transforms</span>
-
- <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">image</span><span class="p">,</span> <span class="n">gt</span><span class="p">):</span>
- <span class="sd">"""Apply a sequence of transformations to ``image``.</span>
-
-<span class="sd"> For each spatially variant transformation, the ground truth ``gt`` is</span>
-<span class="sd"> transformed respectively.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> image : `numpy.ndarray`</span>
-<span class="sd"> The input image.</span>
-<span class="sd"> gt : `numpy.ndarray`</span>
-<span class="sd"> The corresponding ground truth of ``image``.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> image : `numpy.ndarray`</span>
-<span class="sd"> The transformed image.</span>
-<span class="sd"> gt : `numpy.ndarray`</span>
-<span class="sd"> The transformed ground truth.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># apply transformations to the input image in specified order</span>
- <span class="k">for</span> <span class="n">t</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">transforms</span><span class="p">:</span>
- <span class="n">image</span> <span class="o">=</span> <span class="n">t</span><span class="p">(</span><span class="n">image</span><span class="p">)</span>
-
- <span class="c1"># check whether the transformations are invariant and if not, apply</span>
- <span class="c1"># the transformation also to the ground truth</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="n">t</span><span class="o">.</span><span class="n">invariant</span> <span class="ow">and</span> <span class="n">t</span><span class="o">.</span><span class="n">applied</span><span class="p">:</span>
- <span class="c1"># get the probability with which the transformation is applied</span>
- <span class="n">p</span> <span class="o">=</span> <span class="n">t</span><span class="o">.</span><span class="n">p</span>
-
- <span class="c1"># overwrite the probability to 1, to ensure the same</span>
- <span class="c1"># transformation is applied on the ground truth</span>
- <span class="n">t</span><span class="o">.</span><span class="n">p</span> <span class="o">=</span> <span class="mi">1</span>
-
- <span class="c1"># transform ground truth</span>
- <span class="n">gt</span> <span class="o">=</span> <span class="n">t</span><span class="p">(</span><span class="n">gt</span><span class="p">)</span>
-
- <span class="c1"># reset probability to original value</span>
- <span class="n">t</span><span class="o">.</span><span class="n">p</span> <span class="o">=</span> <span class="n">p</span>
-
- <span class="k">return</span> <span class="n">image</span><span class="p">,</span> <span class="n">gt</span>
-
- <span class="k">def</span> <span class="fm">__repr__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Representation of `~pysegcnn.core.transforms.Augment`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> repr : `str`</span>
-<span class="sd"> Representation string.</span>
-
-<span class="sd"> """</span>
- <span class="n">fstring</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="vm">__class__</span><span class="o">.</span><span class="vm">__name__</span> <span class="o">+</span> <span class="s1">'('</span>
- <span class="k">for</span> <span class="n">t</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">transforms</span><span class="p">:</span>
- <span class="n">fstring</span> <span class="o">+=</span> <span class="s1">'</span><span class="se">\n</span><span class="s1">'</span>
- <span class="n">fstring</span> <span class="o">+=</span> <span class="s1">' </span><span class="si">{0}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">t</span><span class="p">)</span>
- <span class="n">fstring</span> <span class="o">+=</span> <span class="s1">'</span><span class="se">\n</span><span class="s1">)'</span>
- <span class="k">return</span> <span class="n">fstring</span></div>
-</pre></div>
-
- </div>
-
- </div>
- <footer>
-
-
- <hr/>
-
- <div role="contentinfo">
- <p>
-
- © Copyright 2020, Daniel Frisinghelli
-
- </p>
- </div>
-
-
-
- Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a
-
- <a href="https://github.com/rtfd/sphinx_rtd_theme">theme</a>
-
- provided by <a href="https://readthedocs.org">Read the Docs</a>.
-
-</footer>
-
- </div>
- </div>
-
- </section>
-
- </div>
-
-
- <script type="text/javascript">
- jQuery(function () {
- SphinxRtdTheme.Navigation.enable(true);
- });
- </script>
-
-
-
-
-
-
-</body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/_modules/core/utils.html b/docs/_build/html/_modules/core/utils.html
deleted file mode 100644
index 7b1803b5f45f93da49115c5fcce7191586238e68..0000000000000000000000000000000000000000
--- a/docs/_build/html/_modules/core/utils.html
+++ /dev/null
@@ -1,1481 +0,0 @@
-
-
-<!DOCTYPE html>
-<html class="writer-html5" lang="en" >
-<head>
- <meta charset="utf-8">
-
- <meta name="viewport" content="width=device-width, initial-scale=1.0">
-
- <title>core.utils — PySegCNN documentation</title>
-
-
-
- <link rel="stylesheet" href="../../_static/css/theme.css" type="text/css" />
- <link rel="stylesheet" href="../../_static/pygments.css" type="text/css" />
-
-
-
-
-
-
-
- <!--[if lt IE 9]>
- <script src="../../_static/js/html5shiv.min.js"></script>
- <![endif]-->
-
-
- <script type="text/javascript" id="documentation_options" data-url_root="../../" src="../../_static/documentation_options.js"></script>
- <script src="../../_static/jquery.js"></script>
- <script src="../../_static/underscore.js"></script>
- <script src="../../_static/doctools.js"></script>
- <script src="../../_static/language_data.js"></script>
-
- <script type="text/javascript" src="../../_static/js/theme.js"></script>
-
-
- <link rel="index" title="Index" href="../../genindex.html" />
- <link rel="search" title="Search" href="../../search.html" />
-</head>
-
-<body class="wy-body-for-nav">
-
-
- <div class="wy-grid-for-nav">
-
- <nav data-toggle="wy-nav-shift" class="wy-nav-side">
- <div class="wy-side-scroll">
- <div class="wy-side-nav-search" >
-
-
-
- <a href="../../index.html" class="icon icon-home" alt="Documentation Home"> PySegCNN
-
-
-
- </a>
-
-
-
-
-
-
-
-<div role="search">
- <form id="rtd-search-form" class="wy-form" action="../../search.html" method="get">
- <input type="text" name="q" placeholder="Search docs" />
- <input type="hidden" name="check_keywords" value="yes" />
- <input type="hidden" name="area" value="default" />
- </form>
-</div>
-
-
- </div>
-
-
- <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
-
-
-
-
-
-
- <ul>
-<li class="toctree-l1"><a class="reference internal" href="../../source/requirements.html">Requirements</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../../source/installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../../source/api.html">API Reference</a></li>
-</ul>
-
-
-
- </div>
-
- </div>
- </nav>
-
- <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
-
-
- <nav class="wy-nav-top" aria-label="top navigation">
-
- <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
- <a href="../../index.html">PySegCNN</a>
-
- </nav>
-
-
- <div class="wy-nav-content">
-
- <div class="rst-content">
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-<div role="navigation" aria-label="breadcrumbs navigation">
-
- <ul class="wy-breadcrumbs">
-
- <li><a href="../../index.html" class="icon icon-home"></a> »</li>
-
- <li><a href="../index.html">Module code</a> »</li>
-
- <li>core.utils</li>
-
-
- <li class="wy-breadcrumbs-aside">
-
- </li>
-
- </ul>
-
-
- <hr/>
-</div>
- <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
- <div itemprop="articleBody">
-
- <h1>Source code for core.utils</h1><div class="highlight"><pre>
-<span></span><span class="sd">"""Utility functions mainly for image IO and reshaping.</span>
-
-<span class="sd">License</span>
-<span class="sd">-------</span>
-
-<span class="sd"> Copyright (c) 2020 Daniel Frisinghelli</span>
-
-<span class="sd"> This source code is licensed under the GNU General Public License v3.</span>
-
-<span class="sd"> See the LICENSE file in the repository's root directory.</span>
-
-<span class="sd">"""</span>
-
-<span class="c1"># !/usr/bin/env python</span>
-<span class="c1"># -*- coding: utf-8 -*-</span>
-
-<span class="c1"># builtins</span>
-<span class="kn">import</span> <span class="nn">os</span>
-<span class="kn">import</span> <span class="nn">re</span>
-<span class="kn">import</span> <span class="nn">shutil</span>
-<span class="kn">import</span> <span class="nn">logging</span>
-<span class="kn">import</span> <span class="nn">pathlib</span>
-<span class="kn">import</span> <span class="nn">tarfile</span>
-<span class="kn">import</span> <span class="nn">zipfile</span>
-<span class="kn">import</span> <span class="nn">datetime</span>
-<span class="kn">import</span> <span class="nn">warnings</span>
-
-<span class="c1"># externals</span>
-<span class="kn">import</span> <span class="nn">gdal</span>
-<span class="kn">import</span> <span class="nn">torch</span>
-<span class="kn">import</span> <span class="nn">numpy</span> <span class="k">as</span> <span class="nn">np</span>
-
-<span class="c1"># module level logger</span>
-<span class="n">LOGGER</span> <span class="o">=</span> <span class="n">logging</span><span class="o">.</span><span class="n">getLogger</span><span class="p">(</span><span class="vm">__name__</span><span class="p">)</span>
-
-<span class="c1"># suffixes for radiometrically calibrated scenes</span>
-<span class="n">SUFFIXES</span> <span class="o">=</span> <span class="p">[</span><span class="s1">'toa_ref'</span><span class="p">,</span> <span class="s1">'toa_rad'</span><span class="p">,</span> <span class="s1">'toa_brt'</span><span class="p">]</span>
-
-
-<div class="viewcode-block" id="img2np"><a class="viewcode-back" href="../../source/core.html#core.utils.img2np">[docs]</a><span class="k">def</span> <span class="nf">img2np</span><span class="p">(</span><span class="n">path</span><span class="p">,</span> <span class="n">tile_size</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">tile</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">pad</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">cval</span><span class="o">=</span><span class="mi">0</span><span class="p">):</span>
- <span class="sd">"""Read an image to a `numpy.ndarray`.</span>
-
-<span class="sd"> If ``tile_size`` is not `None`, the input image is divided into square</span>
-<span class="sd"> tiles of size (``tile_size``, ``tile_size``). If the image is not evenly</span>
-<span class="sd"> divisible and ``pad`` = False, a `ValueError` is raised. However, if</span>
-<span class="sd"> ``pad`` = True, center padding with constant value ``cval`` is applied.</span>
-
-<span class="sd"> The tiling works as follows:</span>
-
-<span class="sd"> (Padded) Input image:</span>
-
-<span class="sd"> ------------------------------------------------</span>
-<span class="sd"> | | | | |</span>
-<span class="sd"> | tile_00 | tile_01 | ... | tile_0n |</span>
-<span class="sd"> | | | | |</span>
-<span class="sd"> |----------------------------------------------|</span>
-<span class="sd"> | | | | |</span>
-<span class="sd"> | tile_10 | tile_11 | ... | tile_1n |</span>
-<span class="sd"> | | | | |</span>
-<span class="sd"> |----------------------------------------------|</span>
-<span class="sd"> | | | | |</span>
-<span class="sd"> | ... | ... | ... | ... |</span>
-<span class="sd"> | | | | |</span>
-<span class="sd"> |----------------------------------------------|</span>
-<span class="sd"> | | | | |</span>
-<span class="sd"> | tile_m0 | tile_m1 | ... | tile_mn |</span>
-<span class="sd"> | | | | |</span>
-<span class="sd"> ------------------------------------------------</span>
-
-<span class="sd"> where m = n. Each tile has its id, which starts at 0 in the topleft corner</span>
-<span class="sd"> of the input image, i.e. tile_00 has id=0, and increases along the width</span>
-<span class="sd"> axis, i.e. tile_0n has id=n, tile_10 has id=n+1, ..., tile_mn has</span>
-<span class="sd"> id=(m * n) - 1.</span>
-
-<span class="sd"> If ``tile`` is an integer, only the tile with id = ``tile`` is returned.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> path : `str` or `None` or `numpy.ndarray`</span>
-<span class="sd"> The image to read.</span>
-<span class="sd"> tile_size : `None` or `int`, optional</span>
-<span class="sd"> The size of a tile. The default is None.</span>
-<span class="sd"> tile : `int`, optional</span>
-<span class="sd"> The tile id. The default is None.</span>
-<span class="sd"> pad : `bool`, optional</span>
-<span class="sd"> Whether to center pad the input image. The default is False.</span>
-<span class="sd"> cval : `float`, optional</span>
-<span class="sd"> The constant padding value. The default is 0.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> FileNotFoundError</span>
-<span class="sd"> Raised if ``path`` is a path that does not exist.</span>
-<span class="sd"> TypeError</span>
-<span class="sd"> Raised if ``path`` is not `str` or `None` or `numpy.ndarray`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> image : `numpy.ndarray`</span>
-<span class="sd"> The image array. The output shape is:</span>
-
-<span class="sd"> if ``tile_size`` is not `None`:</span>
-<span class="sd"> shape=(tiles, bands, tile_size, tile_size)</span>
-<span class="sd"> if the image does only have one band:</span>
-<span class="sd"> shape=(tiles, tile_size, tile_size)</span>
-
-<span class="sd"> else:</span>
-<span class="sd"> shape=(bands, height, width)</span>
-<span class="sd"> if the image does only have one band:</span>
-<span class="sd"> shape=(height, width)</span>
-
-<span class="sd"> """</span>
- <span class="c1"># check the type of path</span>
- <span class="k">if</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">path</span><span class="p">,</span> <span class="nb">str</span><span class="p">):</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">exists</span><span class="p">(</span><span class="n">path</span><span class="p">):</span>
- <span class="k">raise</span> <span class="ne">FileNotFoundError</span><span class="p">(</span><span class="s1">'</span><span class="si">{}</span><span class="s1"> does not exist.'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">path</span><span class="p">))</span>
- <span class="c1"># the image dataset as returned by gdal</span>
- <span class="n">img</span> <span class="o">=</span> <span class="n">gdal</span><span class="o">.</span><span class="n">Open</span><span class="p">(</span><span class="n">path</span><span class="p">)</span>
-
- <span class="c1"># number of bands</span>
- <span class="n">bands</span> <span class="o">=</span> <span class="n">img</span><span class="o">.</span><span class="n">RasterCount</span>
-
- <span class="c1"># spatial size</span>
- <span class="n">height</span> <span class="o">=</span> <span class="n">img</span><span class="o">.</span><span class="n">RasterYSize</span>
- <span class="n">width</span> <span class="o">=</span> <span class="n">img</span><span class="o">.</span><span class="n">RasterXSize</span>
-
- <span class="k">elif</span> <span class="n">path</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">warning</span><span class="p">(</span><span class="s1">'Path is of NoneType, returning.'</span><span class="p">)</span>
- <span class="k">return</span>
-
- <span class="c1"># accept numpy arrays as input</span>
- <span class="k">elif</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">path</span><span class="p">,</span> <span class="n">np</span><span class="o">.</span><span class="n">ndarray</span><span class="p">):</span>
- <span class="n">img</span> <span class="o">=</span> <span class="n">path</span>
- <span class="n">bands</span> <span class="o">=</span> <span class="n">img</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span>
- <span class="n">height</span> <span class="o">=</span> <span class="n">img</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span>
- <span class="n">width</span> <span class="o">=</span> <span class="n">img</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">2</span><span class="p">]</span>
-
- <span class="k">else</span><span class="p">:</span>
- <span class="k">raise</span> <span class="ne">TypeError</span><span class="p">(</span><span class="s1">'Input of type </span><span class="si">{}</span><span class="s1"> not supported'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="nb">type</span><span class="p">(</span><span class="n">img</span><span class="p">)))</span>
-
- <span class="c1"># check whether to read the image in tiles</span>
- <span class="k">if</span> <span class="n">tile_size</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
-
- <span class="c1"># number of tiles</span>
- <span class="n">ntiles</span> <span class="o">=</span> <span class="mi">1</span>
-
- <span class="c1"># create empty numpy array to store whole image</span>
- <span class="n">image</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">empty</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="p">(</span><span class="n">ntiles</span><span class="p">,</span> <span class="n">bands</span><span class="p">,</span> <span class="n">height</span><span class="p">,</span> <span class="n">width</span><span class="p">))</span>
-
- <span class="c1"># iterate over the bands of the image</span>
- <span class="k">for</span> <span class="n">b</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">bands</span><span class="p">):</span>
-
- <span class="c1"># read the data of band b</span>
- <span class="k">if</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">img</span><span class="p">,</span> <span class="n">np</span><span class="o">.</span><span class="n">ndarray</span><span class="p">):</span>
- <span class="n">data</span> <span class="o">=</span> <span class="n">img</span><span class="p">[</span><span class="n">b</span><span class="p">,</span> <span class="o">...</span><span class="p">]</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="n">band</span> <span class="o">=</span> <span class="n">img</span><span class="o">.</span><span class="n">GetRasterBand</span><span class="p">(</span><span class="n">b</span><span class="o">+</span><span class="mi">1</span><span class="p">)</span>
- <span class="n">data</span> <span class="o">=</span> <span class="n">band</span><span class="o">.</span><span class="n">ReadAsArray</span><span class="p">()</span>
-
- <span class="c1"># append band b to numpy image array</span>
- <span class="n">image</span><span class="p">[</span><span class="mi">0</span><span class="p">,</span> <span class="n">b</span><span class="p">,</span> <span class="p">:,</span> <span class="p">:]</span> <span class="o">=</span> <span class="n">data</span>
-
- <span class="k">else</span><span class="p">:</span>
-
- <span class="c1"># check whether the image is evenly divisible in square tiles</span>
- <span class="c1"># of size (tile_size x tile_size)</span>
- <span class="n">ntiles</span><span class="p">,</span> <span class="n">padding</span> <span class="o">=</span> <span class="n">is_divisible</span><span class="p">((</span><span class="n">height</span><span class="p">,</span> <span class="n">width</span><span class="p">),</span> <span class="n">tile_size</span><span class="p">,</span> <span class="n">pad</span><span class="p">)</span>
-
- <span class="c1"># image size after padding</span>
- <span class="n">y_size</span> <span class="o">=</span> <span class="n">height</span> <span class="o">+</span> <span class="n">padding</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">+</span> <span class="n">padding</span><span class="p">[</span><span class="mi">2</span><span class="p">]</span>
- <span class="n">x_size</span> <span class="o">=</span> <span class="n">width</span> <span class="o">+</span> <span class="n">padding</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">+</span> <span class="n">padding</span><span class="p">[</span><span class="mi">3</span><span class="p">]</span>
-
- <span class="c1"># print progress</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">debug</span><span class="p">(</span><span class="s1">'Image size: </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">((</span><span class="n">height</span><span class="p">,</span> <span class="n">width</span><span class="p">)))</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">debug</span><span class="p">(</span><span class="s1">'Dividing image into </span><span class="si">{}</span><span class="s1"> tiles of size </span><span class="si">{}</span><span class="s1"> ...'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">ntiles</span><span class="p">,</span> <span class="p">(</span><span class="n">tile_size</span><span class="p">,</span> <span class="n">tile_size</span><span class="p">)))</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">debug</span><span class="p">(</span><span class="s1">'Padding image (b, l, t, r): </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="nb">tuple</span><span class="p">(</span><span class="n">padding</span><span class="p">)))</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">debug</span><span class="p">(</span><span class="s1">'Padded image size: </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">((</span><span class="n">y_size</span><span class="p">,</span> <span class="n">x_size</span><span class="p">)))</span>
-
- <span class="c1"># get the indices of the top left corner for each tile</span>
- <span class="n">topleft</span> <span class="o">=</span> <span class="n">tile_topleft_corner</span><span class="p">((</span><span class="n">y_size</span><span class="p">,</span> <span class="n">x_size</span><span class="p">),</span> <span class="n">tile_size</span><span class="p">)</span>
-
- <span class="c1"># whether to read all tiles or a single tile</span>
- <span class="k">if</span> <span class="n">tile</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
- <span class="n">ntiles</span> <span class="o">=</span> <span class="mi">1</span>
-
- <span class="c1"># create empty numpy array to store the tiles</span>
- <span class="n">image</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">ones</span><span class="p">((</span><span class="n">ntiles</span><span class="p">,</span> <span class="n">bands</span><span class="p">,</span> <span class="n">tile_size</span><span class="p">,</span> <span class="n">tile_size</span><span class="p">))</span> <span class="o">*</span> <span class="n">cval</span>
-
- <span class="c1"># iterate over the topleft corners of the tiles</span>
- <span class="k">for</span> <span class="n">k</span><span class="p">,</span> <span class="n">corner</span> <span class="ow">in</span> <span class="n">topleft</span><span class="o">.</span><span class="n">items</span><span class="p">():</span>
-
- <span class="c1"># in case a single tile is required, skip the rest of the tiles</span>
- <span class="k">if</span> <span class="n">tile</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
- <span class="k">if</span> <span class="n">k</span> <span class="o">!=</span> <span class="n">tile</span><span class="p">:</span>
- <span class="k">continue</span>
-
- <span class="c1"># set the key to 0 for correct array indexing when reading</span>
- <span class="c1"># a single tile from the image</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">debug</span><span class="p">(</span><span class="s1">'Processing tile </span><span class="si">{}</span><span class="s1"> ...'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">k</span><span class="p">))</span>
- <span class="n">k</span> <span class="o">=</span> <span class="mi">0</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">debug</span><span class="p">(</span><span class="s1">'Creating tile </span><span class="si">{}</span><span class="s1"> with top-left corner </span><span class="si">{}</span><span class="s1"> ...'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">k</span><span class="p">,</span> <span class="n">corner</span><span class="p">))</span>
-
- <span class="c1"># calculate shift between padded and original image</span>
- <span class="n">row</span> <span class="o">=</span> <span class="n">corner</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">-</span> <span class="n">padding</span><span class="p">[</span><span class="mi">2</span><span class="p">]</span> <span class="k">if</span> <span class="n">corner</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">></span> <span class="mi">0</span> <span class="k">else</span> <span class="n">corner</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span>
- <span class="n">col</span> <span class="o">=</span> <span class="n">corner</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">-</span> <span class="n">padding</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="k">if</span> <span class="n">corner</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">></span> <span class="mi">0</span> <span class="k">else</span> <span class="n">corner</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span>
- <span class="n">y_tl</span> <span class="o">=</span> <span class="n">row</span> <span class="o">+</span> <span class="n">padding</span><span class="p">[</span><span class="mi">2</span><span class="p">]</span> <span class="k">if</span> <span class="n">row</span> <span class="o">==</span> <span class="mi">0</span> <span class="k">else</span> <span class="mi">0</span>
- <span class="n">x_tl</span> <span class="o">=</span> <span class="n">col</span> <span class="o">+</span> <span class="n">padding</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="k">if</span> <span class="n">col</span> <span class="o">==</span> <span class="mi">0</span> <span class="k">else</span> <span class="mi">0</span>
-
- <span class="c1"># iterate over the bands of the image</span>
- <span class="k">for</span> <span class="n">b</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">bands</span><span class="p">):</span>
-
- <span class="c1"># check if the current tile extend exists in the image</span>
- <span class="n">nrows</span><span class="p">,</span> <span class="n">ncols</span> <span class="o">=</span> <span class="n">check_tile_extend</span><span class="p">(</span>
- <span class="p">(</span><span class="n">height</span><span class="p">,</span> <span class="n">width</span><span class="p">),</span> <span class="p">(</span><span class="n">row</span><span class="p">,</span> <span class="n">col</span><span class="p">),</span> <span class="n">tile_size</span><span class="p">)</span>
-
- <span class="c1"># read the current tile from band b</span>
- <span class="k">if</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">img</span><span class="p">,</span> <span class="n">np</span><span class="o">.</span><span class="n">ndarray</span><span class="p">):</span>
- <span class="n">data</span> <span class="o">=</span> <span class="n">img</span><span class="p">[</span><span class="n">b</span><span class="p">,</span> <span class="n">row</span><span class="p">:</span><span class="n">row</span><span class="o">+</span><span class="n">nrows</span><span class="p">,</span> <span class="n">col</span><span class="p">:</span><span class="n">col</span><span class="o">+</span><span class="n">ncols</span><span class="p">]</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="n">band</span> <span class="o">=</span> <span class="n">img</span><span class="o">.</span><span class="n">GetRasterBand</span><span class="p">(</span><span class="n">b</span><span class="o">+</span><span class="mi">1</span><span class="p">)</span>
- <span class="n">data</span> <span class="o">=</span> <span class="n">band</span><span class="o">.</span><span class="n">ReadAsArray</span><span class="p">(</span><span class="n">col</span><span class="p">,</span> <span class="n">row</span><span class="p">,</span> <span class="n">ncols</span><span class="p">,</span> <span class="n">nrows</span><span class="p">)</span>
-
- <span class="c1"># append band b to numpy image array</span>
- <span class="n">image</span><span class="p">[</span><span class="n">k</span><span class="p">,</span> <span class="n">b</span><span class="p">,</span> <span class="n">y_tl</span><span class="p">:</span><span class="n">nrows</span><span class="p">,</span> <span class="n">x_tl</span><span class="p">:</span><span class="n">ncols</span><span class="p">]</span> <span class="o">=</span> <span class="n">data</span><span class="p">[</span><span class="mi">0</span><span class="p">:(</span><span class="n">nrows</span> <span class="o">-</span> <span class="n">y_tl</span><span class="p">),</span>
- <span class="mi">0</span><span class="p">:(</span><span class="n">ncols</span> <span class="o">-</span> <span class="n">x_tl</span><span class="p">)]</span>
-
- <span class="c1"># check if there are more than 1 band</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="n">bands</span> <span class="o">></span> <span class="mi">1</span><span class="p">:</span>
- <span class="n">image</span> <span class="o">=</span> <span class="n">image</span><span class="o">.</span><span class="n">squeeze</span><span class="p">(</span><span class="n">axis</span><span class="o">=</span><span class="mi">1</span><span class="p">)</span>
-
- <span class="c1"># check if there are more than 1 tile</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="n">ntiles</span> <span class="o">></span> <span class="mi">1</span><span class="p">:</span>
- <span class="n">image</span> <span class="o">=</span> <span class="n">image</span><span class="o">.</span><span class="n">squeeze</span><span class="p">(</span><span class="n">axis</span><span class="o">=</span><span class="mi">0</span><span class="p">)</span>
-
- <span class="c1"># close tif file</span>
- <span class="k">del</span> <span class="n">img</span>
-
- <span class="c1"># return the image</span>
- <span class="k">return</span> <span class="n">image</span></div>
-
-
-<div class="viewcode-block" id="is_divisible"><a class="viewcode-back" href="../../source/core.html#core.utils.is_divisible">[docs]</a><span class="k">def</span> <span class="nf">is_divisible</span><span class="p">(</span><span class="n">img_size</span><span class="p">,</span> <span class="n">tile_size</span><span class="p">,</span> <span class="n">pad</span><span class="o">=</span><span class="kc">False</span><span class="p">):</span>
- <span class="sd">"""Check whether an image is evenly divisible into square tiles.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> img_size : `tuple`</span>
-<span class="sd"> The image size (height, width).</span>
-<span class="sd"> tile_size : `int`</span>
-<span class="sd"> The size of the tile.</span>
-<span class="sd"> pad : `bool`, optional</span>
-<span class="sd"> Whether to center pad the input image. The default is False.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> ValueError</span>
-<span class="sd"> Raised if the image is not evenly divisible and ``pad`` = False.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> ntiles : `int`</span>
-<span class="sd"> The number of tiles fitting ``img_size``.</span>
-<span class="sd"> padding : `tuple`</span>
-<span class="sd"> The amount of padding (bottom, left, top, right).</span>
-
-<span class="sd"> """</span>
- <span class="c1"># calculate number of pixels per tile</span>
- <span class="n">pixels_per_tile</span> <span class="o">=</span> <span class="n">tile_size</span> <span class="o">**</span> <span class="mi">2</span>
-
- <span class="c1"># check whether the image is evenly divisible in square tiles of size</span>
- <span class="c1"># (tile_size, tile_size)</span>
- <span class="n">ntiles</span> <span class="o">=</span> <span class="p">((</span><span class="n">img_size</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">*</span> <span class="n">img_size</span><span class="p">[</span><span class="mi">1</span><span class="p">])</span> <span class="o">/</span> <span class="n">pixels_per_tile</span><span class="p">)</span>
-
- <span class="c1"># if it is evenly divisible, no padding is required</span>
- <span class="k">if</span> <span class="n">ntiles</span><span class="o">.</span><span class="n">is_integer</span><span class="p">():</span>
- <span class="n">padding</span> <span class="o">=</span> <span class="mi">4</span> <span class="o">*</span> <span class="p">(</span><span class="mi">0</span><span class="p">,)</span>
-
- <span class="k">if</span> <span class="ow">not</span> <span class="n">ntiles</span><span class="o">.</span><span class="n">is_integer</span><span class="p">()</span> <span class="ow">and</span> <span class="ow">not</span> <span class="n">pad</span><span class="p">:</span>
- <span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="s1">'Image of size </span><span class="si">{}</span><span class="s1"> not evenly divisible in (</span><span class="si">{}</span><span class="s1">, </span><span class="si">{}</span><span class="s1">) '</span>
- <span class="s1">'tiles.'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">img_size</span><span class="p">,</span> <span class="n">tile_size</span><span class="p">,</span> <span class="n">tile_size</span><span class="p">))</span>
-
- <span class="k">if</span> <span class="ow">not</span> <span class="n">ntiles</span><span class="o">.</span><span class="n">is_integer</span><span class="p">()</span> <span class="ow">and</span> <span class="n">pad</span><span class="p">:</span>
-
- <span class="c1"># calculate the desired image size, i.e. the smallest size that is</span>
- <span class="c1"># evenly divisible into square tiles of size (tile_size, tile_size)</span>
- <span class="n">h_new</span> <span class="o">=</span> <span class="nb">int</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">ceil</span><span class="p">(</span><span class="n">img_size</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">/</span> <span class="n">tile_size</span><span class="p">)</span> <span class="o">*</span> <span class="n">tile_size</span><span class="p">)</span>
- <span class="n">w_new</span> <span class="o">=</span> <span class="nb">int</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">ceil</span><span class="p">(</span><span class="n">img_size</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">/</span> <span class="n">tile_size</span><span class="p">)</span> <span class="o">*</span> <span class="n">tile_size</span><span class="p">)</span>
-
- <span class="c1"># calculate center offset</span>
- <span class="n">dh</span> <span class="o">=</span> <span class="n">h_new</span> <span class="o">-</span> <span class="n">img_size</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span>
- <span class="n">dw</span> <span class="o">=</span> <span class="n">w_new</span> <span class="o">-</span> <span class="n">img_size</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span>
-
- <span class="c1"># check whether the center offsets are even or odd</span>
-
- <span class="c1"># in case both offsets are even, the padding is symmetric on both the</span>
- <span class="c1"># bottom/top and left/right</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="n">dh</span> <span class="o">%</span> <span class="mi">2</span> <span class="ow">and</span> <span class="ow">not</span> <span class="n">dw</span> <span class="o">%</span> <span class="mi">2</span><span class="p">:</span>
- <span class="n">padding</span> <span class="o">=</span> <span class="p">(</span><span class="n">dh</span> <span class="o">//</span> <span class="mi">2</span><span class="p">,</span> <span class="n">dw</span> <span class="o">//</span> <span class="mi">2</span><span class="p">,</span> <span class="n">dh</span> <span class="o">//</span> <span class="mi">2</span><span class="p">,</span> <span class="n">dw</span> <span class="o">//</span> <span class="mi">2</span><span class="p">)</span>
-
- <span class="c1"># in case only one offset is even, the padding is symmetric along the</span>
- <span class="c1"># even offset and asymmetric along the odd offset</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="n">dh</span> <span class="o">%</span> <span class="mi">2</span> <span class="ow">and</span> <span class="n">dw</span> <span class="o">%</span> <span class="mi">2</span><span class="p">:</span>
- <span class="n">padding</span> <span class="o">=</span> <span class="p">(</span><span class="n">dh</span> <span class="o">//</span> <span class="mi">2</span><span class="p">,</span> <span class="n">dw</span> <span class="o">//</span> <span class="mi">2</span><span class="p">,</span> <span class="n">dh</span> <span class="o">//</span> <span class="mi">2</span><span class="p">,</span> <span class="n">dw</span> <span class="o">//</span> <span class="mi">2</span> <span class="o">+</span> <span class="mi">1</span><span class="p">)</span>
- <span class="k">if</span> <span class="n">dh</span> <span class="o">%</span> <span class="mi">2</span> <span class="ow">and</span> <span class="ow">not</span> <span class="n">dw</span> <span class="o">%</span> <span class="mi">2</span><span class="p">:</span>
- <span class="n">padding</span> <span class="o">=</span> <span class="p">(</span><span class="n">dh</span> <span class="o">//</span> <span class="mi">2</span><span class="p">,</span> <span class="n">dw</span> <span class="o">//</span> <span class="mi">2</span><span class="p">,</span> <span class="n">dh</span> <span class="o">//</span> <span class="mi">2</span> <span class="o">+</span> <span class="mi">1</span><span class="p">,</span> <span class="n">dw</span> <span class="o">//</span> <span class="mi">2</span><span class="p">)</span>
-
- <span class="c1"># in case of offsets are odd, the padding is asymmetric on both the</span>
- <span class="c1"># bottom/top and left/right</span>
- <span class="k">if</span> <span class="n">dh</span> <span class="o">%</span> <span class="mi">2</span> <span class="ow">and</span> <span class="n">dw</span> <span class="o">%</span> <span class="mi">2</span><span class="p">:</span>
- <span class="n">padding</span> <span class="o">=</span> <span class="p">(</span><span class="n">dh</span> <span class="o">//</span> <span class="mi">2</span><span class="p">,</span> <span class="n">dw</span> <span class="o">//</span> <span class="mi">2</span><span class="p">,</span> <span class="n">dh</span> <span class="o">//</span> <span class="mi">2</span> <span class="o">+</span> <span class="mi">1</span><span class="p">,</span> <span class="n">dw</span> <span class="o">//</span> <span class="mi">2</span> <span class="o">+</span> <span class="mi">1</span><span class="p">)</span>
-
- <span class="c1"># calculate number of tiles on padded image</span>
- <span class="n">ntiles</span> <span class="o">=</span> <span class="p">(</span><span class="n">h_new</span> <span class="o">*</span> <span class="n">w_new</span><span class="p">)</span> <span class="o">/</span> <span class="p">(</span><span class="n">tile_size</span> <span class="o">**</span> <span class="mi">2</span><span class="p">)</span>
-
- <span class="k">return</span> <span class="nb">int</span><span class="p">(</span><span class="n">ntiles</span><span class="p">),</span> <span class="n">padding</span></div>
-
-
-<div class="viewcode-block" id="check_tile_extend"><a class="viewcode-back" href="../../source/core.html#core.utils.check_tile_extend">[docs]</a><span class="k">def</span> <span class="nf">check_tile_extend</span><span class="p">(</span><span class="n">img_size</span><span class="p">,</span> <span class="n">topleft</span><span class="p">,</span> <span class="n">tile_size</span><span class="p">):</span>
- <span class="sd">"""Check if a tile exceeds the image size.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> img_size : `tuple`</span>
-<span class="sd"> The image size (height, width).</span>
-<span class="sd"> topleft : `tuple`</span>
-<span class="sd"> The topleft corner of the tile (y, x).</span>
-<span class="sd"> tile_size : `int`</span>
-<span class="sd"> The size of the tile.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> nrows : `int`</span>
-<span class="sd"> Number of rows of the tile within the image.</span>
-<span class="sd"> ncols : TYPE</span>
-<span class="sd"> Number of columns of the tile within the image.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># check if the tile is within both the rows and the columns of the image</span>
- <span class="k">if</span> <span class="p">(</span><span class="n">topleft</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">+</span> <span class="n">tile_size</span> <span class="o"><</span> <span class="n">img_size</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="ow">and</span>
- <span class="n">topleft</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">+</span> <span class="n">tile_size</span> <span class="o"><</span> <span class="n">img_size</span><span class="p">[</span><span class="mi">1</span><span class="p">]):</span>
-
- <span class="c1"># both rows and columns can be equal to the tile size</span>
- <span class="n">nrows</span><span class="p">,</span> <span class="n">ncols</span> <span class="o">=</span> <span class="n">tile_size</span><span class="p">,</span> <span class="n">tile_size</span>
-
- <span class="c1"># check if the tile exceeds one of rows or columns of the image</span>
- <span class="k">if</span> <span class="p">(</span><span class="n">topleft</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">+</span> <span class="n">tile_size</span> <span class="o"><</span> <span class="n">img_size</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="ow">and</span> <span class="ow">not</span>
- <span class="n">topleft</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">+</span> <span class="n">tile_size</span> <span class="o"><</span> <span class="n">img_size</span><span class="p">[</span><span class="mi">1</span><span class="p">]):</span>
-
- <span class="c1"># adjust columns to remaining columns in the original image</span>
- <span class="n">nrows</span><span class="p">,</span> <span class="n">ncols</span> <span class="o">=</span> <span class="n">tile_size</span><span class="p">,</span> <span class="n">img_size</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">-</span> <span class="n">topleft</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span>
-
- <span class="k">if</span> <span class="p">(</span><span class="n">topleft</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">+</span> <span class="n">tile_size</span> <span class="o"><</span> <span class="n">img_size</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="ow">and</span> <span class="ow">not</span>
- <span class="n">topleft</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">+</span> <span class="n">tile_size</span> <span class="o"><</span> <span class="n">img_size</span><span class="p">[</span><span class="mi">0</span><span class="p">]):</span>
-
- <span class="c1"># adjust rows to remaining rows in the original image</span>
- <span class="n">nrows</span><span class="p">,</span> <span class="n">ncols</span> <span class="o">=</span> <span class="n">img_size</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">-</span> <span class="n">topleft</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="n">tile_size</span>
-
- <span class="c1"># check if the tile exceeds both the rows and the columns of the image</span>
- <span class="k">if</span> <span class="p">(</span><span class="ow">not</span> <span class="n">topleft</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">+</span> <span class="n">tile_size</span> <span class="o"><</span> <span class="n">img_size</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="ow">and</span> <span class="ow">not</span>
- <span class="n">topleft</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">+</span> <span class="n">tile_size</span> <span class="o"><</span> <span class="n">img_size</span><span class="p">[</span><span class="mi">1</span><span class="p">]):</span>
-
- <span class="c1"># adjust both rows and columns to the remaining ones</span>
- <span class="n">nrows</span><span class="p">,</span> <span class="n">ncols</span> <span class="o">=</span> <span class="n">img_size</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">-</span> <span class="n">topleft</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="n">img_size</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">-</span> <span class="n">topleft</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span>
-
- <span class="k">return</span> <span class="n">nrows</span><span class="p">,</span> <span class="n">ncols</span></div>
-
-
-<div class="viewcode-block" id="tile_topleft_corner"><a class="viewcode-back" href="../../source/core.html#core.utils.tile_topleft_corner">[docs]</a><span class="k">def</span> <span class="nf">tile_topleft_corner</span><span class="p">(</span><span class="n">img_size</span><span class="p">,</span> <span class="n">tile_size</span><span class="p">):</span>
- <span class="sd">"""Return the topleft corners of the tiles in the image.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> img_size : `tuple`</span>
-<span class="sd"> The image size (height, width).</span>
-<span class="sd"> tile_size : `int`</span>
-<span class="sd"> The size of the tile.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> indices : `dict`</span>
-<span class="sd"> The keys of ``indices`` are the tile ids (`int`) and the values are the</span>
-<span class="sd"> topleft corners (`tuple` = (y, x)) of the tiles.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># check if the image is divisible into square tiles of size</span>
- <span class="c1"># (tile_size, tile_size)</span>
- <span class="n">_</span><span class="p">,</span> <span class="n">_</span> <span class="o">=</span> <span class="n">is_divisible</span><span class="p">(</span><span class="n">img_size</span><span class="p">,</span> <span class="n">tile_size</span><span class="p">,</span> <span class="n">pad</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
-
- <span class="c1"># number of tiles along the width (columns) of the image</span>
- <span class="n">ntiles_columns</span> <span class="o">=</span> <span class="nb">int</span><span class="p">(</span><span class="n">img_size</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">/</span> <span class="n">tile_size</span><span class="p">)</span>
-
- <span class="c1"># number of tiles along the height (rows) of the image</span>
- <span class="n">ntiles_rows</span> <span class="o">=</span> <span class="nb">int</span><span class="p">(</span><span class="n">img_size</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">/</span> <span class="n">tile_size</span><span class="p">)</span>
-
- <span class="c1"># get the indices of the top left corner for each tile</span>
- <span class="n">indices</span> <span class="o">=</span> <span class="p">{}</span>
- <span class="n">k</span> <span class="o">=</span> <span class="mi">0</span>
- <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">ntiles_rows</span><span class="p">):</span>
- <span class="k">for</span> <span class="n">j</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">ntiles_columns</span><span class="p">):</span>
- <span class="n">indices</span><span class="p">[</span><span class="n">k</span><span class="p">]</span> <span class="o">=</span> <span class="p">(</span><span class="n">i</span> <span class="o">*</span> <span class="n">tile_size</span><span class="p">,</span> <span class="n">j</span> <span class="o">*</span> <span class="n">tile_size</span><span class="p">)</span>
- <span class="n">k</span> <span class="o">+=</span> <span class="mi">1</span>
-
- <span class="k">return</span> <span class="n">indices</span></div>
-
-
-<div class="viewcode-block" id="reconstruct_scene"><a class="viewcode-back" href="../../source/core.html#core.utils.reconstruct_scene">[docs]</a><span class="k">def</span> <span class="nf">reconstruct_scene</span><span class="p">(</span><span class="n">tiles</span><span class="p">):</span>
- <span class="sd">"""Reconstruct a tiled image.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> tiles : `torch.Tensor` or `numpy.ndarray`</span>
-<span class="sd"> The tiled image, shape=(tiles, bands, tile_size, tile_size) or</span>
-<span class="sd"> shape=(tiles, tile_size, tile_size).</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> image : `numpy.ndarray`</span>
-<span class="sd"> The reconstructed image.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># convert to numpy array</span>
- <span class="n">tiles</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">tiles</span><span class="p">)</span>
-
- <span class="c1"># check the size</span>
- <span class="k">if</span> <span class="n">tiles</span><span class="o">.</span><span class="n">ndim</span> <span class="o">></span> <span class="mi">3</span><span class="p">:</span>
- <span class="n">nbands</span> <span class="o">=</span> <span class="n">tiles</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span>
- <span class="n">tile_size</span> <span class="o">=</span> <span class="n">tiles</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">2</span><span class="p">]</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="n">nbands</span> <span class="o">=</span> <span class="mi">1</span>
- <span class="n">tile_size</span> <span class="o">=</span> <span class="n">tiles</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span>
-
- <span class="c1"># calculate image size</span>
- <span class="n">img_size</span> <span class="o">=</span> <span class="mi">2</span> <span class="o">*</span> <span class="p">(</span><span class="nb">int</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">sqrt</span><span class="p">(</span><span class="n">tiles</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">0</span><span class="p">])</span> <span class="o">*</span> <span class="n">tile_size</span><span class="p">),)</span>
-
- <span class="c1"># calculate the topleft corners of the tiles</span>
- <span class="n">topleft</span> <span class="o">=</span> <span class="n">tile_topleft_corner</span><span class="p">(</span><span class="n">img_size</span><span class="p">,</span> <span class="n">tile_size</span><span class="p">)</span>
-
- <span class="c1"># iterate over the tiles</span>
- <span class="n">scene</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="p">(</span><span class="n">nbands</span><span class="p">,)</span> <span class="o">+</span> <span class="n">img_size</span><span class="p">)</span>
- <span class="k">for</span> <span class="n">t</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">tiles</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">0</span><span class="p">]):</span>
- <span class="n">scene</span><span class="p">[</span><span class="o">...</span><span class="p">,</span>
- <span class="n">topleft</span><span class="p">[</span><span class="n">t</span><span class="p">][</span><span class="mi">0</span><span class="p">]:</span> <span class="n">topleft</span><span class="p">[</span><span class="n">t</span><span class="p">][</span><span class="mi">0</span><span class="p">]</span> <span class="o">+</span> <span class="n">tile_size</span><span class="p">,</span>
- <span class="n">topleft</span><span class="p">[</span><span class="n">t</span><span class="p">][</span><span class="mi">1</span><span class="p">]:</span> <span class="n">topleft</span><span class="p">[</span><span class="n">t</span><span class="p">][</span><span class="mi">1</span><span class="p">]</span> <span class="o">+</span> <span class="n">tile_size</span><span class="p">]</span> <span class="o">=</span> <span class="n">tiles</span><span class="p">[</span><span class="n">t</span><span class="p">,</span> <span class="o">...</span><span class="p">]</span>
-
- <span class="k">return</span> <span class="n">scene</span><span class="o">.</span><span class="n">squeeze</span><span class="p">()</span></div>
-
-
-<div class="viewcode-block" id="accuracy_function"><a class="viewcode-back" href="../../source/core.html#core.utils.accuracy_function">[docs]</a><span class="k">def</span> <span class="nf">accuracy_function</span><span class="p">(</span><span class="n">outputs</span><span class="p">,</span> <span class="n">labels</span><span class="p">):</span>
- <span class="sd">"""Calculate prediction accuracy.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> outputs : `torch.Tensor` or array_like</span>
-<span class="sd"> The model prediction.</span>
-<span class="sd"> labels : `torch.Tensor` or array_like</span>
-<span class="sd"> The ground truth.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> accuracy : `float`</span>
-<span class="sd"> Mean prediction accuracy.</span>
-
-<span class="sd"> """</span>
- <span class="k">if</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">outputs</span><span class="p">,</span> <span class="n">torch</span><span class="o">.</span><span class="n">Tensor</span><span class="p">):</span>
- <span class="k">return</span> <span class="p">(</span><span class="n">outputs</span> <span class="o">==</span> <span class="n">labels</span><span class="p">)</span><span class="o">.</span><span class="n">float</span><span class="p">()</span><span class="o">.</span><span class="n">mean</span><span class="p">()</span><span class="o">.</span><span class="n">item</span><span class="p">()</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="k">return</span> <span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">outputs</span><span class="p">)</span> <span class="o">==</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">labels</span><span class="p">))</span><span class="o">.</span><span class="n">mean</span><span class="p">()</span><span class="o">.</span><span class="n">item</span><span class="p">()</span></div>
-
-
-<div class="viewcode-block" id="parse_landsat_scene"><a class="viewcode-back" href="../../source/core.html#core.utils.parse_landsat_scene">[docs]</a><span class="k">def</span> <span class="nf">parse_landsat_scene</span><span class="p">(</span><span class="n">scene_id</span><span class="p">):</span>
- <span class="sd">"""Parse a Landsat scene identifier.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> scene_id : `str`</span>
-<span class="sd"> A Landsat scene identifier.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> scene : `dict` or `None`</span>
-<span class="sd"> A dictionary containing scene metadata. If `None`, ``scene_id`` is not</span>
-<span class="sd"> a valid Landsat scene identifier.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># Landsat Collection 1 naming convention in regular expression</span>
- <span class="n">sensor</span> <span class="o">=</span> <span class="s1">'L[COTEM]0[1-8]_'</span>
- <span class="n">level</span> <span class="o">=</span> <span class="s1">'L[0-9][A-Z][A-Z]_'</span>
- <span class="n">swath</span> <span class="o">=</span> <span class="s1">'[0-2][0-9]</span><span class="si">{2}</span><span class="s1">[0-1][0-9]</span><span class="si">{2}</span><span class="s1">_'</span>
- <span class="n">date</span> <span class="o">=</span> <span class="s1">'[0-9]</span><span class="si">{4}</span><span class="s1">[0-1][0-9][0-3][0-9]_'</span>
- <span class="n">doy</span> <span class="o">=</span> <span class="s1">'[0-9]</span><span class="si">{4}</span><span class="s1">[0-3][0-9]</span><span class="si">{2}</span><span class="s1">'</span>
- <span class="n">collection</span> <span class="o">=</span> <span class="s1">'0[0-9]_'</span>
- <span class="n">category</span> <span class="o">=</span> <span class="s1">'[A-Z]([A-Z]|[0-9])'</span>
-
- <span class="c1"># Landsat Collection 1 naming</span>
- <span class="n">C1</span> <span class="o">=</span> <span class="p">(</span><span class="n">sensor</span> <span class="o">+</span> <span class="n">level</span> <span class="o">+</span> <span class="n">swath</span> <span class="o">+</span> <span class="n">date</span> <span class="o">+</span> <span class="n">date</span> <span class="o">+</span> <span class="n">collection</span> <span class="o">+</span> <span class="n">category</span><span class="p">)</span>
- <span class="n">Landsat_C1</span> <span class="o">=</span> <span class="n">re</span><span class="o">.</span><span class="n">compile</span><span class="p">(</span><span class="n">C1</span><span class="p">)</span>
-
- <span class="c1"># Landsat naming convention before Collections</span>
- <span class="n">C0</span> <span class="o">=</span> <span class="p">(</span><span class="n">sensor</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="s1">'_'</span><span class="p">,</span> <span class="s1">''</span><span class="p">)</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="s1">'0'</span><span class="p">,</span> <span class="s1">''</span><span class="p">)</span> <span class="o">+</span> <span class="n">swath</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="s1">'_'</span><span class="p">,</span> <span class="s1">''</span><span class="p">)</span> <span class="o">+</span>
- <span class="n">doy</span> <span class="o">+</span> <span class="s1">'[A-Z]</span><span class="si">{3}</span><span class="s1">'</span> <span class="o">+</span> <span class="s1">'[0-9]</span><span class="si">{2}</span><span class="s1">'</span><span class="p">)</span>
- <span class="n">Landsat_C0</span> <span class="o">=</span> <span class="n">re</span><span class="o">.</span><span class="n">compile</span><span class="p">(</span><span class="n">C0</span><span class="p">)</span>
-
- <span class="c1"># mapping from sensor id to sensors</span>
- <span class="n">lsensors</span> <span class="o">=</span> <span class="p">{</span><span class="s1">'E'</span><span class="p">:</span> <span class="s1">'Enhanced Thematic Mapper Plus'</span><span class="p">,</span>
- <span class="s1">'T'</span><span class="p">:</span> <span class="s1">'Thematic Mapper'</span><span class="p">,</span>
- <span class="s1">'M'</span><span class="p">:</span> <span class="s1">'Multispectral Scanner'</span><span class="p">}</span>
- <span class="n">l8sensors</span> <span class="o">=</span> <span class="p">{</span><span class="s1">'C'</span><span class="p">:</span> <span class="s1">'Operational Land Imager (OLI) & Thermal Infrared Sensor'</span>
- <span class="s1">' (TIRS)'</span><span class="p">,</span>
- <span class="s1">'O'</span><span class="p">:</span> <span class="s1">'Operational Land Imager (OLI)'</span><span class="p">,</span>
- <span class="s1">'T'</span><span class="p">:</span> <span class="s1">'Thermal Infrared Sensor (TIRS)'</span><span class="p">,</span>
- <span class="p">}</span>
-
- <span class="c1"># whether a scene identifier matches the Landsat naming convention</span>
- <span class="n">scene</span> <span class="o">=</span> <span class="p">{}</span>
- <span class="k">if</span> <span class="n">Landsat_C0</span><span class="o">.</span><span class="n">search</span><span class="p">(</span><span class="n">scene_id</span><span class="p">):</span>
-
- <span class="c1"># the match of the regular expression</span>
- <span class="n">match</span> <span class="o">=</span> <span class="n">Landsat_C0</span><span class="o">.</span><span class="n">search</span><span class="p">(</span><span class="n">scene_id</span><span class="p">)[</span><span class="mi">0</span><span class="p">]</span>
-
- <span class="c1"># the metadata of the scene identifier</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'id'</span><span class="p">]</span> <span class="o">=</span> <span class="n">match</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'satellite'</span><span class="p">]</span> <span class="o">=</span> <span class="s1">'Landsat </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">match</span><span class="p">[</span><span class="mi">2</span><span class="p">])</span>
- <span class="k">if</span> <span class="nb">int</span><span class="p">(</span><span class="n">match</span><span class="p">[</span><span class="mi">2</span><span class="p">])</span> <span class="o">></span> <span class="mi">7</span><span class="p">:</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'sensor'</span><span class="p">]</span> <span class="o">=</span> <span class="n">l8sensors</span><span class="p">[</span><span class="n">match</span><span class="p">[</span><span class="mi">1</span><span class="p">]]</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'sensor'</span><span class="p">]</span> <span class="o">=</span> <span class="n">lsensors</span><span class="p">[</span><span class="n">match</span><span class="p">[</span><span class="mi">1</span><span class="p">]]</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'path'</span><span class="p">]</span> <span class="o">=</span> <span class="n">match</span><span class="p">[</span><span class="mi">3</span><span class="p">:</span><span class="mi">6</span><span class="p">]</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'row'</span><span class="p">]</span> <span class="o">=</span> <span class="n">match</span><span class="p">[</span><span class="mi">6</span><span class="p">:</span><span class="mi">9</span><span class="p">]</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'date'</span><span class="p">]</span> <span class="o">=</span> <span class="n">doy2date</span><span class="p">(</span><span class="n">match</span><span class="p">[</span><span class="mi">9</span><span class="p">:</span><span class="mi">13</span><span class="p">],</span> <span class="n">match</span><span class="p">[</span><span class="mi">13</span><span class="p">:</span><span class="mi">16</span><span class="p">])</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'gsi'</span><span class="p">]</span> <span class="o">=</span> <span class="n">match</span><span class="p">[</span><span class="mi">16</span><span class="p">:</span><span class="mi">19</span><span class="p">]</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'version'</span><span class="p">]</span> <span class="o">=</span> <span class="n">match</span><span class="p">[</span><span class="mi">19</span><span class="p">:]</span>
-
- <span class="k">elif</span> <span class="n">Landsat_C1</span><span class="o">.</span><span class="n">search</span><span class="p">(</span><span class="n">scene_id</span><span class="p">):</span>
-
- <span class="c1"># the match of the regular expression</span>
- <span class="n">match</span> <span class="o">=</span> <span class="n">Landsat_C1</span><span class="o">.</span><span class="n">search</span><span class="p">(</span><span class="n">scene_id</span><span class="p">)[</span><span class="mi">0</span><span class="p">]</span>
-
- <span class="c1"># split scene into respective parts</span>
- <span class="n">parts</span> <span class="o">=</span> <span class="n">match</span><span class="o">.</span><span class="n">split</span><span class="p">(</span><span class="s1">'_'</span><span class="p">)</span>
-
- <span class="c1"># the metadata of the scene identifier</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'id'</span><span class="p">]</span> <span class="o">=</span> <span class="n">match</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'satellite'</span><span class="p">]</span> <span class="o">=</span> <span class="s1">'Landsat </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">parts</span><span class="p">[</span><span class="mi">0</span><span class="p">][</span><span class="mi">2</span><span class="p">:])</span>
- <span class="k">if</span> <span class="nb">int</span><span class="p">(</span><span class="n">parts</span><span class="p">[</span><span class="mi">0</span><span class="p">][</span><span class="mi">3</span><span class="p">])</span> <span class="o">></span> <span class="mi">7</span><span class="p">:</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'sensor'</span><span class="p">]</span> <span class="o">=</span> <span class="n">l8sensors</span><span class="p">[</span><span class="n">parts</span><span class="p">[</span><span class="mi">0</span><span class="p">][</span><span class="mi">1</span><span class="p">]]</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'sensor'</span><span class="p">]</span> <span class="o">=</span> <span class="n">lsensors</span><span class="p">[</span><span class="n">parts</span><span class="p">[</span><span class="mi">0</span><span class="p">][</span><span class="mi">1</span><span class="p">]]</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'path'</span><span class="p">]</span> <span class="o">=</span> <span class="n">parts</span><span class="p">[</span><span class="mi">2</span><span class="p">][</span><span class="mi">0</span><span class="p">:</span><span class="mi">3</span><span class="p">]</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'row'</span><span class="p">]</span> <span class="o">=</span> <span class="n">parts</span><span class="p">[</span><span class="mi">2</span><span class="p">][</span><span class="mi">3</span><span class="p">:]</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'date'</span><span class="p">]</span> <span class="o">=</span> <span class="n">datetime</span><span class="o">.</span><span class="n">datetime</span><span class="o">.</span><span class="n">strptime</span><span class="p">(</span><span class="n">parts</span><span class="p">[</span><span class="mi">3</span><span class="p">],</span> <span class="s1">'%Y%m</span><span class="si">%d</span><span class="s1">'</span><span class="p">)</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'collection'</span><span class="p">]</span> <span class="o">=</span> <span class="nb">int</span><span class="p">(</span><span class="n">parts</span><span class="p">[</span><span class="mi">5</span><span class="p">])</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'version'</span><span class="p">]</span> <span class="o">=</span> <span class="n">parts</span><span class="p">[</span><span class="mi">6</span><span class="p">]</span>
-
- <span class="k">else</span><span class="p">:</span>
- <span class="n">scene</span> <span class="o">=</span> <span class="kc">None</span>
-
- <span class="k">return</span> <span class="n">scene</span></div>
-
-
-<div class="viewcode-block" id="parse_sentinel2_scene"><a class="viewcode-back" href="../../source/core.html#core.utils.parse_sentinel2_scene">[docs]</a><span class="k">def</span> <span class="nf">parse_sentinel2_scene</span><span class="p">(</span><span class="n">scene_id</span><span class="p">):</span>
- <span class="sd">"""Parse a Sentinel-2 scene identifier.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> scene_id : `str`</span>
-<span class="sd"> A Sentinel-2 scene identifier.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> scene : `dict` or `None`</span>
-<span class="sd"> A dictionary containing scene metadata. If `None`, ``scene_id`` is not</span>
-<span class="sd"> a valid Sentinel-2 scene identifier.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># Sentinel 2 Level-1C products naming convention after 6th December 2016</span>
- <span class="n">mission</span> <span class="o">=</span> <span class="s1">'S2[A-B]_'</span>
- <span class="n">level</span> <span class="o">=</span> <span class="s1">'MSIL1C_'</span>
- <span class="n">date</span> <span class="o">=</span> <span class="s1">'[0-9]</span><span class="si">{4}</span><span class="s1">[0-1][0-9][0-3][0-9]'</span>
- <span class="n">time</span> <span class="o">=</span> <span class="s1">'T[0-2][0-9][0-5][0-9][0-5][0-9]_'</span>
- <span class="n">processing</span> <span class="o">=</span> <span class="s1">'N[0-9]</span><span class="si">{4}</span><span class="s1">_'</span>
- <span class="n">orbit</span> <span class="o">=</span> <span class="s1">'R[0-1][0-9]</span><span class="si">{2}</span><span class="s1">_'</span>
- <span class="n">tile</span> <span class="o">=</span> <span class="s1">'T[0-9]</span><span class="si">{2}</span><span class="s1">[A-Z]</span><span class="si">{3}</span><span class="s1">_'</span>
- <span class="n">level_1C</span> <span class="o">=</span> <span class="p">(</span><span class="n">mission</span> <span class="o">+</span> <span class="n">level</span> <span class="o">+</span> <span class="n">date</span> <span class="o">+</span> <span class="n">time</span> <span class="o">+</span> <span class="n">processing</span> <span class="o">+</span>
- <span class="n">orbit</span> <span class="o">+</span> <span class="n">tile</span> <span class="o">+</span> <span class="n">date</span> <span class="o">+</span> <span class="n">time</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="s1">'_'</span><span class="p">,</span> <span class="s1">''</span><span class="p">))</span>
- <span class="n">S2_L1C_New</span> <span class="o">=</span> <span class="n">re</span><span class="o">.</span><span class="n">compile</span><span class="p">(</span><span class="n">level_1C</span><span class="p">)</span>
-
- <span class="c1"># Sentinel 2 Level-1C products naming convention before 6th December 2016</span>
- <span class="n">file_class</span> <span class="o">=</span> <span class="s1">'[A-Z]</span><span class="si">{4}</span><span class="s1">_'</span>
- <span class="n">file_category</span> <span class="o">=</span> <span class="s1">'[A-Z]</span><span class="si">{3}</span><span class="s1">_'</span>
- <span class="n">file_semantic</span> <span class="o">=</span> <span class="s1">'L[0-1]([ABC]|_)_[A-Z]</span><span class="si">{2}</span><span class="s1">_'</span>
- <span class="n">site</span> <span class="o">=</span> <span class="s1">'[A-Z_]</span><span class="si">{4}</span><span class="s1">_'</span>
- <span class="n">aorbit</span> <span class="o">=</span> <span class="s1">'A[0-9]</span><span class="si">{6}</span><span class="s1">_'</span>
- <span class="n">S2_L1C_Old</span> <span class="o">=</span> <span class="n">re</span><span class="o">.</span><span class="n">compile</span><span class="p">(</span><span class="n">mission</span> <span class="o">+</span> <span class="n">file_class</span> <span class="o">+</span> <span class="n">file_category</span> <span class="o">+</span>
- <span class="n">file_semantic</span> <span class="o">+</span> <span class="n">site</span> <span class="o">+</span> <span class="s1">'V'</span> <span class="o">+</span> <span class="n">date</span> <span class="o">+</span> <span class="n">time</span> <span class="o">+</span> <span class="n">aorbit</span> <span class="o">+</span>
- <span class="n">tile</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="s1">'_'</span><span class="p">,</span> <span class="s1">''</span><span class="p">))</span>
-
- <span class="c1"># ProSnow project naming convention</span>
- <span class="n">ProSnow</span> <span class="o">=</span> <span class="n">re</span><span class="o">.</span><span class="n">compile</span><span class="p">(</span><span class="n">tile</span> <span class="o">+</span> <span class="n">date</span> <span class="o">+</span> <span class="n">time</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="s1">'_'</span><span class="p">,</span> <span class="s1">''</span><span class="p">))</span>
-
- <span class="c1"># whether a scene identifier matches the Sentinel naming convention</span>
- <span class="n">scene</span> <span class="o">=</span> <span class="p">{}</span>
- <span class="k">if</span> <span class="n">S2_L1C_Old</span><span class="o">.</span><span class="n">search</span><span class="p">(</span><span class="n">scene_id</span><span class="p">):</span>
-
- <span class="c1"># the match of the regular expression</span>
- <span class="n">match</span> <span class="o">=</span> <span class="n">S2_L1C_Old</span><span class="o">.</span><span class="n">search</span><span class="p">(</span><span class="n">scene_id</span><span class="p">)[</span><span class="mi">0</span><span class="p">]</span>
-
- <span class="c1"># split scene into respective parts</span>
- <span class="n">parts</span> <span class="o">=</span> <span class="n">match</span><span class="o">.</span><span class="n">split</span><span class="p">(</span><span class="s1">'_'</span><span class="p">)</span>
-
- <span class="c1"># the metadata of the scene identifier</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'id'</span><span class="p">]</span> <span class="o">=</span> <span class="n">match</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'satellite'</span><span class="p">]</span> <span class="o">=</span> <span class="s1">'Sentinel </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">parts</span><span class="p">[</span><span class="mi">1</span><span class="p">:])</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'file class'</span><span class="p">]</span> <span class="o">=</span> <span class="n">parts</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'file category'</span><span class="p">]</span> <span class="o">=</span> <span class="n">parts</span><span class="p">[</span><span class="mi">2</span><span class="p">]</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'file semantic'</span><span class="p">]</span> <span class="o">=</span> <span class="n">parts</span><span class="p">[</span><span class="mi">3</span><span class="p">]</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'site'</span><span class="p">]</span> <span class="o">=</span> <span class="n">parts</span><span class="p">[</span><span class="mi">4</span><span class="p">]</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'orbit'</span><span class="p">]</span> <span class="o">=</span> <span class="n">parts</span><span class="p">[</span><span class="mi">6</span><span class="p">]</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'date'</span><span class="p">]</span> <span class="o">=</span> <span class="n">datetime</span><span class="o">.</span><span class="n">datetime</span><span class="o">.</span><span class="n">strptime</span><span class="p">(</span>
- <span class="n">parts</span><span class="p">[</span><span class="mi">7</span><span class="p">]</span><span class="o">.</span><span class="n">split</span><span class="p">(</span><span class="s1">'T'</span><span class="p">)[</span><span class="mi">0</span><span class="p">]</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="s1">'V'</span><span class="p">,</span> <span class="s1">''</span><span class="p">),</span> <span class="s1">'%Y%m</span><span class="si">%d</span><span class="s1">'</span><span class="p">)</span>
-
- <span class="k">elif</span> <span class="n">S2_L1C_New</span><span class="o">.</span><span class="n">search</span><span class="p">(</span><span class="n">scene_id</span><span class="p">):</span>
-
- <span class="c1"># the match of the regular expression</span>
- <span class="n">match</span> <span class="o">=</span> <span class="n">S2_L1C_New</span><span class="o">.</span><span class="n">search</span><span class="p">(</span><span class="n">scene_id</span><span class="p">)[</span><span class="mi">0</span><span class="p">]</span>
-
- <span class="c1"># split scene into respective parts</span>
- <span class="n">parts</span> <span class="o">=</span> <span class="n">match</span><span class="o">.</span><span class="n">split</span><span class="p">(</span><span class="s1">'_'</span><span class="p">)</span>
-
- <span class="c1"># the metadata of the scene identifier</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'id'</span><span class="p">]</span> <span class="o">=</span> <span class="n">match</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'satellite'</span><span class="p">]</span> <span class="o">=</span> <span class="s1">'Sentinel </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">parts</span><span class="p">[</span><span class="mi">1</span><span class="p">:])</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'product'</span><span class="p">]</span> <span class="o">=</span> <span class="n">parts</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'date'</span><span class="p">]</span> <span class="o">=</span> <span class="n">datetime</span><span class="o">.</span><span class="n">datetime</span><span class="o">.</span><span class="n">strptime</span><span class="p">(</span><span class="n">parts</span><span class="p">[</span><span class="mi">2</span><span class="p">]</span><span class="o">.</span><span class="n">split</span><span class="p">(</span><span class="s1">'T'</span><span class="p">)[</span><span class="mi">0</span><span class="p">],</span>
- <span class="s1">'%Y%m</span><span class="si">%d</span><span class="s1">'</span><span class="p">)</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'baseline'</span><span class="p">]</span> <span class="o">=</span> <span class="n">parts</span><span class="p">[</span><span class="mi">3</span><span class="p">]</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'orbit'</span><span class="p">]</span> <span class="o">=</span> <span class="n">parts</span><span class="p">[</span><span class="mi">4</span><span class="p">]</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'tile'</span><span class="p">]</span> <span class="o">=</span> <span class="n">parts</span><span class="p">[</span><span class="mi">5</span><span class="p">]</span>
-
- <span class="k">elif</span> <span class="n">ProSnow</span><span class="o">.</span><span class="n">search</span><span class="p">(</span><span class="n">scene_id</span><span class="p">):</span>
-
- <span class="c1"># the match of the regular expression</span>
- <span class="n">match</span> <span class="o">=</span> <span class="n">ProSnow</span><span class="o">.</span><span class="n">search</span><span class="p">(</span><span class="n">scene_id</span><span class="p">)[</span><span class="mi">0</span><span class="p">]</span>
-
- <span class="c1"># split scene into respective parts</span>
- <span class="n">parts</span> <span class="o">=</span> <span class="n">match</span><span class="o">.</span><span class="n">split</span><span class="p">(</span><span class="s1">'_'</span><span class="p">)</span>
-
- <span class="c1"># the metadata of the scene identifier</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'id'</span><span class="p">]</span> <span class="o">=</span> <span class="n">match</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'satellite'</span><span class="p">]</span> <span class="o">=</span> <span class="s1">'Sentinel 2'</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'tile'</span><span class="p">]</span> <span class="o">=</span> <span class="n">parts</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span>
- <span class="n">scene</span><span class="p">[</span><span class="s1">'date'</span><span class="p">]</span> <span class="o">=</span> <span class="n">datetime</span><span class="o">.</span><span class="n">datetime</span><span class="o">.</span><span class="n">strptime</span><span class="p">(</span><span class="n">parts</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span><span class="o">.</span><span class="n">split</span><span class="p">(</span><span class="s1">'T'</span><span class="p">)[</span><span class="mi">0</span><span class="p">],</span>
- <span class="s1">'%Y%m</span><span class="si">%d</span><span class="s1">'</span><span class="p">)</span>
-
- <span class="k">else</span><span class="p">:</span>
- <span class="n">scene</span> <span class="o">=</span> <span class="kc">None</span>
-
- <span class="k">return</span> <span class="n">scene</span></div>
-
-
-<div class="viewcode-block" id="doy2date"><a class="viewcode-back" href="../../source/core.html#core.utils.doy2date">[docs]</a><span class="k">def</span> <span class="nf">doy2date</span><span class="p">(</span><span class="n">year</span><span class="p">,</span> <span class="n">doy</span><span class="p">):</span>
- <span class="sd">"""Convert the (year, day of the year) date format to a datetime object.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> year : `int`</span>
-<span class="sd"> The year</span>
-<span class="sd"> doy : `int`</span>
-<span class="sd"> The day of the year</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> date : `datetime.datetime`</span>
-<span class="sd"> The converted date.</span>
-<span class="sd"> """</span>
- <span class="c1"># convert year/day of year to a datetime object</span>
- <span class="n">date</span> <span class="o">=</span> <span class="p">(</span><span class="n">datetime</span><span class="o">.</span><span class="n">datetime</span><span class="p">(</span><span class="nb">int</span><span class="p">(</span><span class="n">year</span><span class="p">),</span> <span class="mi">1</span><span class="p">,</span> <span class="mi">1</span><span class="p">)</span> <span class="o">+</span>
- <span class="n">datetime</span><span class="o">.</span><span class="n">timedelta</span><span class="p">(</span><span class="n">days</span><span class="o">=</span><span class="p">(</span><span class="nb">int</span><span class="p">(</span><span class="n">doy</span><span class="p">)</span> <span class="o">-</span> <span class="mi">1</span><span class="p">)))</span>
-
- <span class="k">return</span> <span class="n">date</span></div>
-
-
-<div class="viewcode-block" id="item_in_enum"><a class="viewcode-back" href="../../source/core.html#core.utils.item_in_enum">[docs]</a><span class="k">def</span> <span class="nf">item_in_enum</span><span class="p">(</span><span class="n">name</span><span class="p">,</span> <span class="n">enum</span><span class="p">):</span>
- <span class="sd">"""Check if an item exists in an enumeration.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> name : `str`</span>
-<span class="sd"> Name of the item.</span>
-<span class="sd"> enum : `enum.Enum`</span>
-<span class="sd"> An instance of `enum.Enum`.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> ValueError</span>
-<span class="sd"> Raised if ``name`` is not in ``enum``.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> value</span>
-<span class="sd"> The value of ``name`` in ``enum``.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># check whether the input name exists in the enumeration</span>
- <span class="k">if</span> <span class="n">name</span> <span class="ow">not</span> <span class="ow">in</span> <span class="n">enum</span><span class="o">.</span><span class="n">__members__</span><span class="p">:</span>
- <span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="s1">'</span><span class="si">{}</span><span class="s1"> is not in </span><span class="si">{}</span><span class="s1"> enumeration. Valid names are: </span><span class="se">\n</span><span class="s1"> </span><span class="si">{}</span><span class="s1">'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">name</span><span class="p">,</span> <span class="n">enum</span><span class="o">.</span><span class="vm">__name__</span><span class="p">,</span>
- <span class="s1">'</span><span class="se">\n</span><span class="s1">'</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="s1">'- </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">n</span><span class="p">)</span> <span class="k">for</span> <span class="n">n</span> <span class="ow">in</span>
- <span class="n">enum</span><span class="o">.</span><span class="n">__members__</span><span class="p">)))</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="k">return</span> <span class="n">enum</span><span class="o">.</span><span class="n">__members__</span><span class="p">[</span><span class="n">name</span><span class="p">]</span><span class="o">.</span><span class="n">value</span></div>
-
-
-<div class="viewcode-block" id="destack_tiff"><a class="viewcode-back" href="../../source/core.html#core.utils.destack_tiff">[docs]</a><span class="k">def</span> <span class="nf">destack_tiff</span><span class="p">(</span><span class="n">image</span><span class="p">,</span> <span class="n">outpath</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">overwrite</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">remove</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span>
- <span class="n">suffix</span><span class="o">=</span><span class="s1">''</span><span class="p">):</span>
- <span class="sd">"""Destack a TIFF with more than one band into a TIFF file for each band.</span>
-
-<span class="sd"> Each band in ``image`` is saved to ``outpath`` as distinct TIFF file.</span>
-<span class="sd"> The default filenames are: "filename(``image``) + _B(i).tif", where i is</span>
-<span class="sd"> the respective number of each band in ``image``.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> image : `str` or `pathlib.Path`</span>
-<span class="sd"> The TIFF to destack.</span>
-<span class="sd"> outpath : `str`, optional</span>
-<span class="sd"> Path to save the output TIFF files. The default is None. If None,</span>
-<span class="sd"> ``outpath`` is the path to ``image``.</span>
-<span class="sd"> remove : `bool`, optional</span>
-<span class="sd"> Whether to remove ``image`` from disk after destacking. The default is</span>
-<span class="sd"> False.</span>
-<span class="sd"> overwrite : `bool`, optional</span>
-<span class="sd"> Whether to overwrite existing TIFF files.</span>
-<span class="sd"> suffix : `str`, optional</span>
-<span class="sd"> String to append to the filename of ``image``. If specified, the TIFF</span>
-<span class="sd"> filenames for each band in ``image`` are, "filename(``image``) +</span>
-<span class="sd"> + _B(i)_ + ``suffix``.tif". The default is ''.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> FileNotFoundError</span>
-<span class="sd"> Raised if ``image`` does not exist.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># stop gdal printing warnings and errors to STDERR</span>
- <span class="n">gdal</span><span class="o">.</span><span class="n">PushErrorHandler</span><span class="p">(</span><span class="s1">'CPLQuietErrorHandler'</span><span class="p">)</span>
-
- <span class="c1"># raise Python exceptions for gdal errors</span>
- <span class="n">gdal</span><span class="o">.</span><span class="n">UseExceptions</span><span class="p">()</span>
-
- <span class="c1"># convert to pathlib.Path</span>
- <span class="n">image</span> <span class="o">=</span> <span class="n">pathlib</span><span class="o">.</span><span class="n">Path</span><span class="p">(</span><span class="n">image</span><span class="p">)</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="n">image</span><span class="o">.</span><span class="n">exists</span><span class="p">():</span>
- <span class="k">raise</span> <span class="ne">FileNotFoundError</span><span class="p">(</span><span class="s1">'</span><span class="si">{}</span><span class="s1"> does not exist.'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">image</span><span class="p">))</span>
-
- <span class="c1"># name of the TIFF</span>
- <span class="n">imgname</span> <span class="o">=</span> <span class="n">image</span><span class="o">.</span><span class="n">stem</span>
-
- <span class="c1"># default: output directory is equal to the input directory</span>
- <span class="k">if</span> <span class="n">outpath</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
- <span class="n">outpath</span> <span class="o">=</span> <span class="n">image</span><span class="o">.</span><span class="n">parent</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="n">outpath</span> <span class="o">=</span> <span class="n">pathlib</span><span class="o">.</span><span class="n">Path</span><span class="p">(</span><span class="n">outpath</span><span class="p">)</span>
- <span class="c1"># check if output directory exists</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="n">outpath</span><span class="o">.</span><span class="n">exists</span><span class="p">():</span>
- <span class="n">outpath</span><span class="o">.</span><span class="n">mkdir</span><span class="p">(</span><span class="n">parents</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span> <span class="n">exist_ok</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
-
- <span class="c1"># open the TIFF</span>
- <span class="n">img</span> <span class="o">=</span> <span class="n">gdal</span><span class="o">.</span><span class="n">Open</span><span class="p">(</span><span class="nb">str</span><span class="p">(</span><span class="n">image</span><span class="p">))</span>
-
- <span class="c1"># check whether the file was already processed</span>
- <span class="n">processed</span> <span class="o">=</span> <span class="nb">list</span><span class="p">(</span><span class="n">outpath</span><span class="o">.</span><span class="n">glob</span><span class="p">(</span><span class="n">imgname</span> <span class="o">+</span> <span class="s1">'_B*'</span><span class="p">))</span>
- <span class="k">if</span> <span class="nb">len</span><span class="p">(</span><span class="n">processed</span><span class="p">)</span> <span class="o">>=</span> <span class="n">img</span><span class="o">.</span><span class="n">RasterCount</span> <span class="ow">and</span> <span class="ow">not</span> <span class="n">overwrite</span><span class="p">:</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'</span><span class="si">{}</span><span class="s1"> already processed.'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">imgname</span><span class="p">))</span>
-
- <span class="c1"># destack the TIFF</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="c1"># image driver</span>
- <span class="n">driver</span> <span class="o">=</span> <span class="n">gdal</span><span class="o">.</span><span class="n">GetDriverByName</span><span class="p">(</span><span class="s1">'GTiff'</span><span class="p">)</span>
- <span class="n">driver</span><span class="o">.</span><span class="n">Register</span><span class="p">()</span>
-
- <span class="c1"># image size and tiles</span>
- <span class="n">cols</span> <span class="o">=</span> <span class="n">img</span><span class="o">.</span><span class="n">RasterXSize</span>
- <span class="n">rows</span> <span class="o">=</span> <span class="n">img</span><span class="o">.</span><span class="n">RasterYSize</span>
- <span class="n">bands</span> <span class="o">=</span> <span class="n">img</span><span class="o">.</span><span class="n">RasterCount</span>
-
- <span class="c1"># print progress</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Processing: </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">imgname</span><span class="p">))</span>
-
- <span class="c1"># iterate the bands of the raster</span>
- <span class="k">for</span> <span class="n">b</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="mi">1</span><span class="p">,</span> <span class="n">bands</span> <span class="o">+</span> <span class="mi">1</span><span class="p">):</span>
-
- <span class="c1"># read the data of band b</span>
- <span class="n">band</span> <span class="o">=</span> <span class="n">img</span><span class="o">.</span><span class="n">GetRasterBand</span><span class="p">(</span><span class="n">b</span><span class="p">)</span>
- <span class="n">data</span> <span class="o">=</span> <span class="n">band</span><span class="o">.</span><span class="n">ReadAsArray</span><span class="p">()</span>
-
- <span class="c1"># output file: replace for band name</span>
- <span class="n">fname</span> <span class="o">=</span> <span class="nb">str</span><span class="p">(</span><span class="n">outpath</span><span class="o">.</span><span class="n">joinpath</span><span class="p">(</span><span class="n">imgname</span> <span class="o">+</span> <span class="s1">'_B</span><span class="si">{:d}</span><span class="s1">.tif'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">b</span><span class="p">)))</span>
- <span class="k">if</span> <span class="n">suffix</span><span class="p">:</span>
- <span class="n">fname</span> <span class="o">=</span> <span class="n">fname</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="s1">'.tif'</span><span class="p">,</span> <span class="s1">'_</span><span class="si">{}</span><span class="s1">.tif'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">suffix</span><span class="p">))</span>
- <span class="n">outDs</span> <span class="o">=</span> <span class="n">driver</span><span class="o">.</span><span class="n">Create</span><span class="p">(</span><span class="n">fname</span><span class="p">,</span> <span class="n">cols</span><span class="p">,</span> <span class="n">rows</span><span class="p">,</span> <span class="mi">1</span><span class="p">,</span> <span class="n">band</span><span class="o">.</span><span class="n">DataType</span><span class="p">)</span>
-
- <span class="c1"># define output band</span>
- <span class="n">outband</span> <span class="o">=</span> <span class="n">outDs</span><span class="o">.</span><span class="n">GetRasterBand</span><span class="p">(</span><span class="mi">1</span><span class="p">)</span>
-
- <span class="c1"># write array to output band</span>
- <span class="n">outband</span><span class="o">.</span><span class="n">WriteArray</span><span class="p">(</span><span class="n">data</span><span class="p">)</span>
- <span class="n">outband</span><span class="o">.</span><span class="n">FlushCache</span><span class="p">()</span>
-
- <span class="c1"># Set the geographic information</span>
- <span class="n">outDs</span><span class="o">.</span><span class="n">SetProjection</span><span class="p">(</span><span class="n">img</span><span class="o">.</span><span class="n">GetProjection</span><span class="p">())</span>
- <span class="n">outDs</span><span class="o">.</span><span class="n">SetGeoTransform</span><span class="p">(</span><span class="n">img</span><span class="o">.</span><span class="n">GetGeoTransform</span><span class="p">())</span>
-
- <span class="c1"># clear memory</span>
- <span class="k">del</span> <span class="n">outband</span><span class="p">,</span> <span class="n">band</span><span class="p">,</span> <span class="n">data</span><span class="p">,</span> <span class="n">outDs</span>
-
- <span class="c1"># remove old stacked GeoTIFF</span>
- <span class="k">del</span> <span class="n">img</span>
- <span class="k">if</span> <span class="n">remove</span><span class="p">:</span>
- <span class="n">os</span><span class="o">.</span><span class="n">unlink</span><span class="p">(</span><span class="n">image</span><span class="p">)</span>
- <span class="k">return</span></div>
-
-
-<div class="viewcode-block" id="standard_eo_structure"><a class="viewcode-back" href="../../source/core.html#core.utils.standard_eo_structure">[docs]</a><span class="k">def</span> <span class="nf">standard_eo_structure</span><span class="p">(</span><span class="n">source_path</span><span class="p">,</span> <span class="n">target_path</span><span class="p">,</span> <span class="n">overwrite</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span>
- <span class="n">move</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">parser</span><span class="o">=</span><span class="n">parse_landsat_scene</span><span class="p">):</span>
- <span class="sd">"""Modify the directory structure of a remote sensing dataset.</span>
-
-<span class="sd"> This function assumes that ``source_path`` points to a directory containing</span>
-<span class="sd"> remote sensing data, where each file in ``source_path`` and its sub-folders</span>
-<span class="sd"> should contain a scene identifier in its filename. The scene identifier is</span>
-<span class="sd"> used to restructure the dataset.</span>
-
-<span class="sd"> Currently, Landsat and Sentinel-2 datasets are supported.</span>
-
-<span class="sd"> The directory tree in ``source_path`` is modified to the following</span>
-<span class="sd"> structure in ``target_path``:</span>
-
-<span class="sd"> target_path/</span>
-<span class="sd"> scene_id_1/</span>
-<span class="sd"> files matching scene_id_1</span>
-<span class="sd"> scene_id_2/</span>
-<span class="sd"> files matching scene_id_2</span>
-<span class="sd"> .</span>
-<span class="sd"> .</span>
-<span class="sd"> .</span>
-<span class="sd"> scene_id_n/</span>
-<span class="sd"> files matching scene_id_n</span>
-
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> source_path : `str` or `pathlib.Path`</span>
-<span class="sd"> Path to the remote sensing dataset.</span>
-<span class="sd"> target_path : `str` or `pathlib.Path`</span>
-<span class="sd"> Path to save the restructured dataset.</span>
-<span class="sd"> overwrite : `bool`, optional</span>
-<span class="sd"> Whether to overwrite existing files in ``target_path``.</span>
-<span class="sd"> The default is True.</span>
-<span class="sd"> move : `bool`, optional</span>
-<span class="sd"> Whether to move the files from ``source_path`` to ``target_path``. If</span>
-<span class="sd"> True, files in ``source_path`` are moved to ``target_path``, if False,</span>
-<span class="sd"> files in ``source_path`` are copied to ``target_path``. The default is</span>
-<span class="sd"> False.</span>
-<span class="sd"> parser : `function`, optional</span>
-<span class="sd"> The scene identifier parsing function. Depends on the sensor of the</span>
-<span class="sd"> dataset. See e.g., `pysegcnn.core.utils.parse_landsat_scene`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
- <span class="c1"># create a directory for each scene</span>
- <span class="k">for</span> <span class="n">dirpath</span><span class="p">,</span> <span class="n">dirnames</span><span class="p">,</span> <span class="n">filenames</span> <span class="ow">in</span> <span class="n">os</span><span class="o">.</span><span class="n">walk</span><span class="p">(</span><span class="n">source_path</span><span class="p">):</span>
-
- <span class="c1"># check if there are files in the current folder</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="n">filenames</span><span class="p">:</span>
- <span class="k">continue</span>
-
- <span class="c1"># iterate over the files to modify</span>
- <span class="k">for</span> <span class="n">file</span> <span class="ow">in</span> <span class="n">filenames</span><span class="p">:</span>
-
- <span class="c1"># get the path to the file</span>
- <span class="n">old_path</span> <span class="o">=</span> <span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">dirpath</span><span class="p">,</span> <span class="n">file</span><span class="p">)</span>
-
- <span class="c1"># get name of the scene</span>
- <span class="n">scene</span> <span class="o">=</span> <span class="n">parser</span><span class="p">(</span><span class="n">old_path</span><span class="p">)</span>
- <span class="k">if</span> <span class="n">scene</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
-
- <span class="c1"># path to copy files not matching a scene identifier</span>
- <span class="n">new_path</span> <span class="o">=</span> <span class="n">pathlib</span><span class="o">.</span><span class="n">Path</span><span class="p">(</span><span class="n">target_path</span><span class="p">)</span><span class="o">.</span><span class="n">joinpath</span><span class="p">(</span><span class="s1">'misc'</span><span class="p">,</span> <span class="n">file</span><span class="p">)</span>
-
- <span class="c1"># file a warning if the file does not match a scene identifier</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">warning</span><span class="p">(</span><span class="s1">'</span><span class="si">{}</span><span class="s1"> does not match a scene identifier. Copying '</span>
- <span class="s1">'to </span><span class="si">{}</span><span class="s1">.'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">basename</span><span class="p">(</span><span class="n">old_path</span><span class="p">),</span>
- <span class="n">new_path</span><span class="o">.</span><span class="n">parent</span><span class="p">))</span>
- <span class="k">else</span><span class="p">:</span>
-
- <span class="c1"># the name of the scene</span>
- <span class="n">scene_name</span> <span class="o">=</span> <span class="n">scene</span><span class="p">[</span><span class="s1">'id'</span><span class="p">]</span>
-
- <span class="c1"># define the new path to the file</span>
- <span class="n">new_path</span> <span class="o">=</span> <span class="n">pathlib</span><span class="o">.</span><span class="n">Path</span><span class="p">(</span><span class="n">target_path</span><span class="p">)</span><span class="o">.</span><span class="n">joinpath</span><span class="p">(</span><span class="n">scene_name</span><span class="p">,</span> <span class="n">file</span><span class="p">)</span>
-
- <span class="c1"># move files to new directory</span>
- <span class="k">if</span> <span class="n">new_path</span><span class="o">.</span><span class="n">is_file</span><span class="p">()</span> <span class="ow">and</span> <span class="ow">not</span> <span class="n">overwrite</span><span class="p">:</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'</span><span class="si">{}</span><span class="s1"> already exists.'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">new_path</span><span class="p">))</span>
- <span class="k">continue</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="n">new_path</span><span class="o">.</span><span class="n">parent</span><span class="o">.</span><span class="n">mkdir</span><span class="p">(</span><span class="n">parents</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span> <span class="n">exist_ok</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
- <span class="k">if</span> <span class="n">move</span><span class="p">:</span>
- <span class="n">shutil</span><span class="o">.</span><span class="n">move</span><span class="p">(</span><span class="n">old_path</span><span class="p">,</span> <span class="n">new_path</span><span class="p">)</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'mv </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">new_path</span><span class="p">))</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'cp </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">new_path</span><span class="p">))</span>
- <span class="n">shutil</span><span class="o">.</span><span class="n">copy</span><span class="p">(</span><span class="n">old_path</span><span class="p">,</span> <span class="n">new_path</span><span class="p">)</span>
-
- <span class="c1"># remove old file location</span>
- <span class="k">if</span> <span class="n">move</span><span class="p">:</span>
- <span class="n">shutil</span><span class="o">.</span><span class="n">rmtree</span><span class="p">(</span><span class="n">source_path</span><span class="p">)</span></div>
-
-
-<div class="viewcode-block" id="extract_archive"><a class="viewcode-back" href="../../source/core.html#core.utils.extract_archive">[docs]</a><span class="k">def</span> <span class="nf">extract_archive</span><span class="p">(</span><span class="n">inpath</span><span class="p">,</span> <span class="n">outpath</span><span class="p">,</span> <span class="n">overwrite</span><span class="o">=</span><span class="kc">False</span><span class="p">):</span>
- <span class="sd">"""Extract files from an archive.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> inpath : `str` or `pathlib.Path`</span>
-<span class="sd"> Path to an archive.</span>
-<span class="sd"> outpath : `str` or `pathlib.Path`</span>
-<span class="sd"> Path to save extracted files.</span>
-<span class="sd"> overwrite : `bool`, optional</span>
-<span class="sd"> Whether to overwrite existing extracted files.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> subdir : str</span>
-<span class="sd"> path to the extracted files</span>
-<span class="sd"> """</span>
- <span class="n">inpath</span> <span class="o">=</span> <span class="n">pathlib</span><span class="o">.</span><span class="n">Path</span><span class="p">(</span><span class="n">inpath</span><span class="p">)</span>
-
- <span class="c1"># create the output directory</span>
- <span class="n">outpath</span> <span class="o">=</span> <span class="n">pathlib</span><span class="o">.</span><span class="n">Path</span><span class="p">(</span><span class="n">outpath</span><span class="p">)</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="n">outpath</span><span class="o">.</span><span class="n">exists</span><span class="p">():</span>
- <span class="n">outpath</span><span class="o">.</span><span class="n">mkdir</span><span class="p">(</span><span class="n">parents</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
-
- <span class="c1"># name of the archive</span>
- <span class="n">archive</span> <span class="o">=</span> <span class="n">inpath</span><span class="o">.</span><span class="n">stem</span>
-
- <span class="c1"># path to the extracted files</span>
- <span class="n">target</span> <span class="o">=</span> <span class="n">outpath</span><span class="o">.</span><span class="n">joinpath</span><span class="p">(</span><span class="n">archive</span><span class="p">)</span>
- <span class="k">if</span> <span class="n">target</span><span class="o">.</span><span class="n">exists</span><span class="p">():</span>
- <span class="k">if</span> <span class="n">overwrite</span><span class="p">:</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Overwriting: </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">target</span><span class="p">))</span>
- <span class="n">shutil</span><span class="o">.</span><span class="n">rmtree</span><span class="p">(</span><span class="n">target</span><span class="p">)</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Extracted files are located in: </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">target</span><span class="p">))</span>
- <span class="k">return</span> <span class="n">target</span>
-
- <span class="c1"># create output directory</span>
- <span class="n">target</span><span class="o">.</span><span class="n">mkdir</span><span class="p">(</span><span class="n">parents</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
-
- <span class="c1"># read the archive</span>
- <span class="k">if</span> <span class="n">inpath</span><span class="o">.</span><span class="n">name</span><span class="o">.</span><span class="n">endswith</span><span class="p">(</span><span class="s1">'.tgz'</span><span class="p">)</span> <span class="ow">or</span> <span class="n">inpath</span><span class="o">.</span><span class="n">name</span><span class="o">.</span><span class="n">endswith</span><span class="p">(</span><span class="s1">'.tar.gz'</span><span class="p">):</span>
- <span class="n">tar</span> <span class="o">=</span> <span class="n">tarfile</span><span class="o">.</span><span class="n">open</span><span class="p">(</span><span class="n">inpath</span><span class="p">,</span> <span class="s2">"r:gz"</span><span class="p">)</span>
- <span class="k">elif</span> <span class="n">inpath</span><span class="o">.</span><span class="n">name</span><span class="o">.</span><span class="n">endswith</span><span class="p">(</span><span class="s1">'.zip'</span><span class="p">):</span>
- <span class="n">tar</span> <span class="o">=</span> <span class="n">zipfile</span><span class="o">.</span><span class="n">ZipFile</span><span class="p">(</span><span class="n">inpath</span><span class="p">,</span> <span class="s1">'r'</span><span class="p">)</span>
-
- <span class="c1"># extract files to the output directory</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Extracting: </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">archive</span><span class="p">))</span>
- <span class="n">tar</span><span class="o">.</span><span class="n">extractall</span><span class="p">(</span><span class="n">target</span><span class="p">)</span>
-
- <span class="k">return</span> <span class="n">target</span></div>
-
-
-<div class="viewcode-block" id="read_landsat_metadata"><a class="viewcode-back" href="../../source/core.html#core.utils.read_landsat_metadata">[docs]</a><span class="k">def</span> <span class="nf">read_landsat_metadata</span><span class="p">(</span><span class="n">file</span><span class="p">):</span>
- <span class="sd">"""Parse the Landsat metadata *_MTL.txt file.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> file : `str` or `pathlib.Path`</span>
-<span class="sd"> Path to a Landsat *_MTL.txt file.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> FileNotFoundError</span>
-<span class="sd"> Raised if ``file`` does not exist.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> metadata : `dict`</span>
-<span class="sd"> The metadata text file as dictionary, where each line is a (key, value)</span>
-<span class="sd"> pair.</span>
-<span class="sd"> """</span>
- <span class="n">file</span> <span class="o">=</span> <span class="n">pathlib</span><span class="o">.</span><span class="n">Path</span><span class="p">(</span><span class="n">file</span><span class="p">)</span>
- <span class="c1"># check if the metadata file exists</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="n">file</span><span class="o">.</span><span class="n">exists</span><span class="p">():</span>
- <span class="k">raise</span> <span class="ne">FileNotFoundError</span><span class="p">(</span><span class="s1">'</span><span class="si">{}</span><span class="s1"> does not exist'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">file</span><span class="p">))</span>
-
- <span class="c1"># read metadata file</span>
- <span class="n">metadata</span> <span class="o">=</span> <span class="p">{}</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Parsing metadata file: </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">file</span><span class="o">.</span><span class="n">name</span><span class="p">))</span>
- <span class="k">with</span> <span class="nb">open</span><span class="p">(</span><span class="n">file</span><span class="p">,</span> <span class="s1">'r'</span><span class="p">)</span> <span class="k">as</span> <span class="n">metafile</span><span class="p">:</span>
- <span class="c1"># iterate over the lines of the metadata file</span>
- <span class="k">for</span> <span class="n">line</span> <span class="ow">in</span> <span class="n">metafile</span><span class="p">:</span>
- <span class="k">try</span><span class="p">:</span>
- <span class="c1"># the current line as (key, value pair)</span>
- <span class="p">(</span><span class="n">key</span><span class="p">,</span> <span class="n">value</span><span class="p">)</span> <span class="o">=</span> <span class="n">line</span><span class="o">.</span><span class="n">split</span><span class="p">(</span><span class="s1">'='</span><span class="p">)</span>
-
- <span class="c1"># store current line in dictionary: skip lines defining the</span>
- <span class="c1"># parameter groups</span>
- <span class="k">if</span> <span class="s1">'GROUP'</span> <span class="ow">not</span> <span class="ow">in</span> <span class="n">key</span><span class="p">:</span>
- <span class="n">metadata</span><span class="p">[</span><span class="n">key</span><span class="o">.</span><span class="n">strip</span><span class="p">()]</span> <span class="o">=</span> <span class="n">value</span><span class="o">.</span><span class="n">strip</span><span class="p">()</span>
-
- <span class="c1"># catch value error of line.split('='), i.e. if there is no '='</span>
- <span class="c1"># sign in the current line</span>
- <span class="k">except</span> <span class="ne">ValueError</span><span class="p">:</span>
- <span class="k">continue</span>
-
- <span class="k">return</span> <span class="n">metadata</span></div>
-
-
-<div class="viewcode-block" id="get_radiometric_constants"><a class="viewcode-back" href="../../source/core.html#core.utils.get_radiometric_constants">[docs]</a><span class="k">def</span> <span class="nf">get_radiometric_constants</span><span class="p">(</span><span class="n">metadata</span><span class="p">):</span>
- <span class="sd">"""Retrieve the radiometric calibration constants.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> metadata : `dict`</span>
-<span class="sd"> The dictionary returned by ``read_landsat_metadata``.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> oli : `dict`</span>
-<span class="sd"> Radiometric rescaling factors of the OLI sensor.</span>
-<span class="sd"> tir : `dict`</span>
-<span class="sd"> Thermal conversion constants of the TIRS sensor.</span>
-<span class="sd"> """</span>
- <span class="c1"># regular expression patterns matching the radiometric rescaling factors</span>
- <span class="n">oli_pattern</span> <span class="o">=</span> <span class="n">re</span><span class="o">.</span><span class="n">compile</span><span class="p">(</span><span class="s1">'(RADIANCE|REFLECTANCE)_(MULT|ADD)_BAND_</span><span class="se">\\</span><span class="s1">d{1,2}'</span><span class="p">)</span>
- <span class="n">tir_pattern</span> <span class="o">=</span> <span class="n">re</span><span class="o">.</span><span class="n">compile</span><span class="p">(</span><span class="s1">'K(1|2)_CONSTANT_BAND_</span><span class="se">\\</span><span class="s1">d</span><span class="si">{2}</span><span class="s1">'</span><span class="p">)</span>
-
- <span class="c1"># rescaling factors to calculate top of atmosphere radiance and reflectance</span>
- <span class="n">oli</span> <span class="o">=</span> <span class="p">{</span><span class="n">key</span><span class="p">:</span> <span class="nb">float</span><span class="p">(</span><span class="n">value</span><span class="p">)</span> <span class="k">for</span> <span class="n">key</span><span class="p">,</span> <span class="n">value</span> <span class="ow">in</span> <span class="n">metadata</span><span class="o">.</span><span class="n">items</span><span class="p">()</span> <span class="k">if</span>
- <span class="n">oli_pattern</span><span class="o">.</span><span class="n">search</span><span class="p">(</span><span class="n">key</span><span class="p">)</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">}</span>
-
- <span class="c1"># rescaling factors to calculate at-satellite temperatures in Kelvin</span>
- <span class="n">tir</span> <span class="o">=</span> <span class="p">{</span><span class="n">key</span><span class="p">:</span> <span class="nb">float</span><span class="p">(</span><span class="n">value</span><span class="p">)</span> <span class="k">for</span> <span class="n">key</span><span class="p">,</span> <span class="n">value</span> <span class="ow">in</span> <span class="n">metadata</span><span class="o">.</span><span class="n">items</span><span class="p">()</span> <span class="k">if</span>
- <span class="n">tir_pattern</span><span class="o">.</span><span class="n">search</span><span class="p">(</span><span class="n">key</span><span class="p">)</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">}</span>
-
- <span class="k">return</span> <span class="n">oli</span><span class="p">,</span> <span class="n">tir</span></div>
-
-
-<div class="viewcode-block" id="landsat_radiometric_calibration"><a class="viewcode-back" href="../../source/core.html#core.utils.landsat_radiometric_calibration">[docs]</a><span class="k">def</span> <span class="nf">landsat_radiometric_calibration</span><span class="p">(</span><span class="n">scene</span><span class="p">,</span> <span class="n">outpath</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">exclude</span><span class="o">=</span><span class="p">[],</span>
- <span class="n">radiance</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">overwrite</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span>
- <span class="n">remove_raw</span><span class="o">=</span><span class="kc">True</span><span class="p">):</span>
- <span class="sd">"""Radiometric calibration of Landsat Collection Level 1 scenes.</span>
-
-<span class="sd"> Convert the Landsat OLI bands to top of atmosphere radiance or reflectance</span>
-<span class="sd"> and the TIRS bands to top of atmosphere brightness temperature.</span>
-
-<span class="sd"> Conversion is performed following the `equations`_ provided by the USGS.</span>
-
-<span class="sd"> The filename of each band is extended by one of the following suffixes,</span>
-<span class="sd"> depending on the type of the radiometric calibration:</span>
-
-<span class="sd"> 'toa_ref': top of atmosphere reflectance</span>
-<span class="sd"> 'toa_rad': top of atmopshere radiance</span>
-<span class="sd"> 'toa_brt': top of atmosphere brightness temperature</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> scene : `str` or `pathlib.Path`</span>
-<span class="sd"> Path to a Landsat scene in digital number format.</span>
-<span class="sd"> outpath : `str` or `pathlib.Path`, optional</span>
-<span class="sd"> Path to save the calibrated images. The default is None, which means</span>
-<span class="sd"> saving to ``scene``.</span>
-<span class="sd"> exclude : `list` [`str`], optional</span>
-<span class="sd"> Bands to exclude from the radiometric calibration. The default is [].</span>
-<span class="sd"> radiance : `bool`, optional</span>
-<span class="sd"> Whether to calculate top of atmosphere radiance. The default is False,</span>
-<span class="sd"> which means calculating top of atmopshere reflectance.</span>
-<span class="sd"> overwrite : `bool`, optional</span>
-<span class="sd"> Whether to overwrite the calibrated images. The default is False.</span>
-<span class="sd"> remove_raw : `bool`, optional</span>
-<span class="sd"> Whether to remove the raw digitial number images. The default is True.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> FileNotFoundError</span>
-<span class="sd"> Raised if ``scene`` does not exist.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> .. _equations:</span>
-<span class="sd"> https://www.usgs.gov/land-resources/nli/landsat/using-usgs-landsat-level-1-data-product</span>
-
-<span class="sd"> """</span>
- <span class="n">scene</span> <span class="o">=</span> <span class="n">pathlib</span><span class="o">.</span><span class="n">Path</span><span class="p">(</span><span class="n">scene</span><span class="p">)</span>
- <span class="c1"># check if the input scene exists</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="n">scene</span><span class="o">.</span><span class="n">exists</span><span class="p">():</span>
- <span class="k">raise</span> <span class="ne">FileNotFoundError</span><span class="p">(</span><span class="s1">'</span><span class="si">{}</span><span class="s1"> does not exist.'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">scene</span><span class="p">))</span>
-
- <span class="c1"># default: output directory is equal to the input directory</span>
- <span class="k">if</span> <span class="n">outpath</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
- <span class="n">outpath</span> <span class="o">=</span> <span class="n">scene</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="n">outpath</span> <span class="o">=</span> <span class="n">pathlib</span><span class="o">.</span><span class="n">Path</span><span class="p">(</span><span class="n">outpath</span><span class="p">)</span>
- <span class="c1"># check if output directory exists</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="n">outpath</span><span class="o">.</span><span class="n">exists</span><span class="p">():</span>
- <span class="n">outpath</span><span class="o">.</span><span class="n">mkdir</span><span class="p">(</span><span class="n">parents</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span> <span class="n">exist_ok</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
-
- <span class="c1"># the scene metadata file</span>
- <span class="k">try</span><span class="p">:</span>
- <span class="n">mpattern</span> <span class="o">=</span> <span class="n">re</span><span class="o">.</span><span class="n">compile</span><span class="p">(</span><span class="s1">'[mMtTlL].txt'</span><span class="p">)</span>
- <span class="n">metafile</span> <span class="o">=</span> <span class="p">[</span><span class="n">f</span> <span class="k">for</span> <span class="n">f</span> <span class="ow">in</span> <span class="n">scene</span><span class="o">.</span><span class="n">iterdir</span><span class="p">()</span> <span class="k">if</span> <span class="n">mpattern</span><span class="o">.</span><span class="n">search</span><span class="p">(</span><span class="nb">str</span><span class="p">(</span><span class="n">f</span><span class="p">))]</span>
- <span class="n">metadata</span> <span class="o">=</span> <span class="n">read_landsat_metadata</span><span class="p">(</span><span class="n">metafile</span><span class="o">.</span><span class="n">pop</span><span class="p">())</span>
- <span class="k">except</span> <span class="ne">IndexError</span><span class="p">:</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">error</span><span class="p">(</span><span class="s1">'Can not calibrate scene </span><span class="si">{}</span><span class="s1">: </span><span class="si">{}</span><span class="s1"> does not exist.'</span>
- <span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">scene</span><span class="o">.</span><span class="n">name</span><span class="p">,</span> <span class="n">scene</span><span class="o">.</span><span class="n">name</span> <span class="o">+</span> <span class="s1">'_MTL.txt'</span><span class="p">))</span>
- <span class="k">return</span>
-
- <span class="c1"># radiometric calibration constants</span>
- <span class="n">oli</span><span class="p">,</span> <span class="n">tir</span> <span class="o">=</span> <span class="n">get_radiometric_constants</span><span class="p">(</span><span class="n">metadata</span><span class="p">)</span>
-
- <span class="c1"># log current Landsat scene ID</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Landsat scene id: </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">metadata</span><span class="p">[</span><span class="s1">'LANDSAT_SCENE_ID'</span><span class="p">]))</span>
-
- <span class="c1"># images to process</span>
- <span class="n">ipattern</span> <span class="o">=</span> <span class="n">re</span><span class="o">.</span><span class="n">compile</span><span class="p">(</span><span class="s1">'B</span><span class="se">\\</span><span class="s1">d{1,2}(.*)</span><span class="se">\\</span><span class="s1">.[tT][iI][fF]'</span><span class="p">)</span>
- <span class="n">images</span> <span class="o">=</span> <span class="p">[</span><span class="n">file</span> <span class="k">for</span> <span class="n">file</span> <span class="ow">in</span> <span class="n">scene</span><span class="o">.</span><span class="n">iterdir</span><span class="p">()</span> <span class="k">if</span> <span class="n">ipattern</span><span class="o">.</span><span class="n">search</span><span class="p">(</span><span class="nb">str</span><span class="p">(</span><span class="n">file</span><span class="p">))]</span>
-
- <span class="c1"># pattern to match calibrated images</span>
- <span class="n">cal_pattern</span> <span class="o">=</span> <span class="n">re</span><span class="o">.</span><span class="n">compile</span><span class="p">(</span><span class="s1">'(</span><span class="si">{}</span><span class="s1">|</span><span class="si">{}</span><span class="s1">|</span><span class="si">{}</span><span class="s1">).[tT][iI][fF]'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="o">*</span><span class="n">SUFFIXES</span><span class="p">))</span>
-
- <span class="c1"># check if any images were already processe</span>
- <span class="n">processed</span> <span class="o">=</span> <span class="p">[</span><span class="n">file</span> <span class="k">for</span> <span class="n">file</span> <span class="ow">in</span> <span class="n">images</span> <span class="k">if</span> <span class="n">cal_pattern</span><span class="o">.</span><span class="n">search</span><span class="p">(</span><span class="nb">str</span><span class="p">(</span><span class="n">file</span><span class="p">))]</span>
- <span class="k">if</span> <span class="nb">any</span><span class="p">(</span><span class="n">processed</span><span class="p">):</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'The following images have already been processed:'</span><span class="p">)</span>
-
- <span class="c1"># images that were already processed</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">((</span><span class="s1">'</span><span class="se">\n</span><span class="s1"> '</span> <span class="o">+</span> <span class="p">(</span><span class="nb">len</span><span class="p">(</span><span class="vm">__name__</span><span class="p">)</span> <span class="o">+</span> <span class="mi">1</span><span class="p">)</span> <span class="o">*</span> <span class="s1">' '</span><span class="p">)</span><span class="o">.</span><span class="n">join</span><span class="p">(</span>
- <span class="p">[</span><span class="nb">str</span><span class="p">(</span><span class="n">file</span><span class="o">.</span><span class="n">name</span><span class="p">)</span> <span class="k">for</span> <span class="n">file</span> <span class="ow">in</span> <span class="n">processed</span><span class="p">]))</span>
-
- <span class="c1"># overwrite: remove processed images and redo calibration</span>
- <span class="k">if</span> <span class="n">overwrite</span><span class="p">:</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Preparing to overwrite ...'</span><span class="p">)</span>
-
- <span class="c1"># remove processed images</span>
- <span class="k">for</span> <span class="n">toremove</span> <span class="ow">in</span> <span class="n">processed</span><span class="p">:</span>
- <span class="c1"># remove from disk</span>
- <span class="n">os</span><span class="o">.</span><span class="n">unlink</span><span class="p">(</span><span class="n">toremove</span><span class="p">)</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'rm </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">toremove</span><span class="p">))</span>
-
- <span class="c1"># remove from list to process</span>
- <span class="n">images</span><span class="o">.</span><span class="n">remove</span><span class="p">(</span><span class="n">toremove</span><span class="p">)</span>
-
- <span class="c1"># not overwriting, terminate calibration</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="k">return</span>
-
- <span class="c1"># exclude defined bands from the calibration procedure</span>
- <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="n">images</span><span class="p">:</span>
- <span class="n">current_band</span> <span class="o">=</span> <span class="n">re</span><span class="o">.</span><span class="n">search</span><span class="p">(</span><span class="s1">'B</span><span class="se">\\</span><span class="s1">d{1,2}'</span><span class="p">,</span> <span class="nb">str</span><span class="p">(</span><span class="n">i</span><span class="p">))[</span><span class="mi">0</span><span class="p">]</span>
- <span class="k">if</span> <span class="n">current_band</span> <span class="ow">in</span> <span class="n">exclude</span><span class="p">:</span>
- <span class="n">images</span><span class="o">.</span><span class="n">remove</span><span class="p">(</span><span class="n">i</span><span class="p">)</span>
-
- <span class="c1"># image driver</span>
- <span class="n">driver</span> <span class="o">=</span> <span class="n">gdal</span><span class="o">.</span><span class="n">GetDriverByName</span><span class="p">(</span><span class="s1">'GTiff'</span><span class="p">)</span>
- <span class="n">driver</span><span class="o">.</span><span class="n">Register</span><span class="p">()</span>
-
- <span class="c1"># iterate over the different bands</span>
- <span class="k">for</span> <span class="n">image</span> <span class="ow">in</span> <span class="n">images</span><span class="p">:</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Processing: </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">image</span><span class="o">.</span><span class="n">name</span><span class="p">))</span>
-
- <span class="c1"># read the image</span>
- <span class="n">img</span> <span class="o">=</span> <span class="n">gdal</span><span class="o">.</span><span class="n">Open</span><span class="p">(</span><span class="nb">str</span><span class="p">(</span><span class="n">image</span><span class="p">))</span>
- <span class="n">band</span> <span class="o">=</span> <span class="n">img</span><span class="o">.</span><span class="n">GetRasterBand</span><span class="p">(</span><span class="mi">1</span><span class="p">)</span>
-
- <span class="c1"># read data as array</span>
- <span class="n">data</span> <span class="o">=</span> <span class="n">band</span><span class="o">.</span><span class="n">ReadAsArray</span><span class="p">()</span>
-
- <span class="c1"># mask of erroneous values, i.e. mask of values < 0</span>
- <span class="n">mask</span> <span class="o">=</span> <span class="n">data</span> <span class="o"><</span> <span class="mi">0</span>
-
- <span class="c1"># output filename</span>
- <span class="n">fname</span> <span class="o">=</span> <span class="n">outpath</span><span class="o">.</span><span class="n">joinpath</span><span class="p">(</span><span class="n">image</span><span class="o">.</span><span class="n">stem</span><span class="p">)</span>
-
- <span class="c1"># get the current band</span>
- <span class="n">band</span> <span class="o">=</span> <span class="n">re</span><span class="o">.</span><span class="n">search</span><span class="p">(</span><span class="s1">'B</span><span class="se">\\</span><span class="s1">d{1,2}'</span><span class="p">,</span> <span class="nb">str</span><span class="p">(</span><span class="n">image</span><span class="p">))[</span><span class="mi">0</span><span class="p">]</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="s1">'B'</span><span class="p">,</span> <span class="s1">'BAND_'</span><span class="p">)</span>
-
- <span class="c1"># check if the current band is a thermal band</span>
- <span class="k">if</span> <span class="n">band</span> <span class="ow">in</span> <span class="p">[</span><span class="s1">'BAND_10'</span><span class="p">,</span> <span class="s1">'BAND_11'</span><span class="p">]:</span>
-
- <span class="c1"># output file name for TIRS bands</span>
- <span class="n">fname</span> <span class="o">=</span> <span class="n">pathlib</span><span class="o">.</span><span class="n">Path</span><span class="p">(</span><span class="nb">str</span><span class="p">(</span><span class="n">fname</span><span class="p">)</span> <span class="o">+</span> <span class="s1">'_toa_brt.tif'</span><span class="p">)</span>
-
- <span class="c1"># calculate top of atmosphere brightness temperature</span>
- <span class="k">with</span> <span class="n">warnings</span><span class="o">.</span><span class="n">catch_warnings</span><span class="p">():</span>
- <span class="n">warnings</span><span class="o">.</span><span class="n">filterwarnings</span><span class="p">(</span><span class="s2">"ignore"</span><span class="p">,</span> <span class="n">category</span><span class="o">=</span><span class="ne">RuntimeWarning</span><span class="p">)</span>
-
- <span class="c1"># top of atmosphere radiance</span>
- <span class="n">rad</span> <span class="o">=</span> <span class="p">(</span><span class="n">oli</span><span class="p">[</span><span class="s1">'RADIANCE_MULT_</span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">band</span><span class="p">)]</span> <span class="o">*</span> <span class="n">data</span> <span class="o">+</span>
- <span class="n">oli</span><span class="p">[</span><span class="s1">'RADIANCE_ADD_</span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">band</span><span class="p">)])</span>
-
- <span class="c1"># top of atmosphere brightness temperature</span>
- <span class="n">den</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">log</span><span class="p">(</span><span class="n">tir</span><span class="p">[</span><span class="s1">'K1_CONSTANT_</span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">band</span><span class="p">)]</span> <span class="o">/</span> <span class="n">rad</span> <span class="o">+</span> <span class="mi">1</span><span class="p">)</span>
- <span class="n">toa</span> <span class="o">=</span> <span class="n">tir</span><span class="p">[</span><span class="s1">'K2_CONSTANT_</span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">band</span><span class="p">)]</span> <span class="o">/</span> <span class="n">den</span>
-
- <span class="c1"># clear memory</span>
- <span class="k">del</span> <span class="n">den</span><span class="p">,</span> <span class="n">data</span><span class="p">,</span> <span class="n">rad</span>
- <span class="k">else</span><span class="p">:</span>
-
- <span class="c1"># whether to calculate top of atmosphere radiance or reflectance</span>
- <span class="k">if</span> <span class="n">radiance</span><span class="p">:</span>
-
- <span class="c1"># output file name for OLI bands: radiance</span>
- <span class="n">fname</span> <span class="o">=</span> <span class="n">pathlib</span><span class="o">.</span><span class="n">Path</span><span class="p">(</span><span class="nb">str</span><span class="p">(</span><span class="n">fname</span><span class="p">)</span> <span class="o">+</span> <span class="s1">'_toa_rad.tif'</span><span class="p">)</span>
-
- <span class="c1"># calculate top of atmosphere radiance</span>
- <span class="n">toa</span> <span class="o">=</span> <span class="p">(</span><span class="n">oli</span><span class="p">[</span><span class="s1">'RADIANCE_MULT_</span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">band</span><span class="p">)]</span> <span class="o">*</span> <span class="n">data</span> <span class="o">+</span>
- <span class="n">oli</span><span class="p">[</span><span class="s1">'RADIANCE_ADD_</span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">band</span><span class="p">)])</span>
-
- <span class="c1"># clear memory</span>
- <span class="k">del</span> <span class="n">data</span>
-
- <span class="k">else</span><span class="p">:</span>
-
- <span class="c1"># output file name for OLI bands: reflectance</span>
- <span class="n">fname</span> <span class="o">=</span> <span class="n">pathlib</span><span class="o">.</span><span class="n">Path</span><span class="p">(</span><span class="nb">str</span><span class="p">(</span><span class="n">fname</span><span class="p">)</span> <span class="o">+</span> <span class="s1">'_toa_ref.tif'</span><span class="p">)</span>
-
- <span class="c1"># solar zenith angle in radians</span>
- <span class="n">zenith</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">radians</span><span class="p">(</span><span class="mi">90</span> <span class="o">-</span> <span class="nb">float</span><span class="p">(</span><span class="n">metadata</span><span class="p">[</span><span class="s1">'SUN_ELEVATION'</span><span class="p">]))</span>
-
- <span class="c1"># calculate top of the atmosphere reflectance</span>
- <span class="n">ref</span> <span class="o">=</span> <span class="p">(</span><span class="n">oli</span><span class="p">[</span><span class="s1">'REFLECTANCE_MULT_</span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">band</span><span class="p">)]</span> <span class="o">*</span> <span class="n">data</span> <span class="o">+</span>
- <span class="n">oli</span><span class="p">[</span><span class="s1">'REFLECTANCE_ADD_</span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">band</span><span class="p">)])</span>
- <span class="n">toa</span> <span class="o">=</span> <span class="n">ref</span> <span class="o">/</span> <span class="n">np</span><span class="o">.</span><span class="n">cos</span><span class="p">(</span><span class="n">zenith</span><span class="p">)</span>
-
- <span class="c1"># clear memory</span>
- <span class="k">del</span> <span class="n">ref</span><span class="p">,</span> <span class="n">data</span>
-
- <span class="c1"># mask erroneous values</span>
- <span class="n">toa</span><span class="p">[</span><span class="n">mask</span><span class="p">]</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">nan</span>
-
- <span class="c1"># output file</span>
- <span class="n">outDs</span> <span class="o">=</span> <span class="n">driver</span><span class="o">.</span><span class="n">Create</span><span class="p">(</span><span class="nb">str</span><span class="p">(</span><span class="n">fname</span><span class="p">),</span> <span class="n">img</span><span class="o">.</span><span class="n">RasterXSize</span><span class="p">,</span> <span class="n">img</span><span class="o">.</span><span class="n">RasterYSize</span><span class="p">,</span>
- <span class="n">img</span><span class="o">.</span><span class="n">RasterCount</span><span class="p">,</span> <span class="n">gdal</span><span class="o">.</span><span class="n">GDT_Float32</span><span class="p">)</span>
- <span class="n">outband</span> <span class="o">=</span> <span class="n">outDs</span><span class="o">.</span><span class="n">GetRasterBand</span><span class="p">(</span><span class="mi">1</span><span class="p">)</span>
-
- <span class="c1"># write array</span>
- <span class="n">outband</span><span class="o">.</span><span class="n">WriteArray</span><span class="p">(</span><span class="n">toa</span><span class="p">)</span>
- <span class="n">outband</span><span class="o">.</span><span class="n">FlushCache</span><span class="p">()</span>
-
- <span class="c1"># Set the geographic information</span>
- <span class="n">outDs</span><span class="o">.</span><span class="n">SetProjection</span><span class="p">(</span><span class="n">img</span><span class="o">.</span><span class="n">GetProjection</span><span class="p">())</span>
- <span class="n">outDs</span><span class="o">.</span><span class="n">SetGeoTransform</span><span class="p">(</span><span class="n">img</span><span class="o">.</span><span class="n">GetGeoTransform</span><span class="p">())</span>
-
- <span class="c1"># clear memory</span>
- <span class="k">del</span> <span class="n">outband</span><span class="p">,</span> <span class="n">band</span><span class="p">,</span> <span class="n">img</span><span class="p">,</span> <span class="n">outDs</span><span class="p">,</span> <span class="n">toa</span><span class="p">,</span> <span class="n">mask</span>
-
- <span class="c1"># check if raw images should be removed</span>
- <span class="k">if</span> <span class="n">remove_raw</span><span class="p">:</span>
-
- <span class="c1"># raw digital number images</span>
- <span class="n">_dn</span> <span class="o">=</span> <span class="p">[</span><span class="n">i</span> <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="n">scene</span><span class="o">.</span><span class="n">iterdir</span><span class="p">()</span> <span class="k">if</span> <span class="n">ipattern</span><span class="o">.</span><span class="n">search</span><span class="p">(</span><span class="nb">str</span><span class="p">(</span><span class="n">i</span><span class="p">))</span> <span class="ow">and</span>
- <span class="ow">not</span> <span class="n">cal_pattern</span><span class="o">.</span><span class="n">search</span><span class="p">(</span><span class="nb">str</span><span class="p">(</span><span class="n">i</span><span class="p">))]</span>
-
- <span class="c1"># remove raw digital number images</span>
- <span class="k">for</span> <span class="n">toremove</span> <span class="ow">in</span> <span class="n">_dn</span><span class="p">:</span>
- <span class="c1"># remove from disk</span>
- <span class="n">os</span><span class="o">.</span><span class="n">unlink</span><span class="p">(</span><span class="n">toremove</span><span class="p">)</span>
- <span class="n">LOGGER</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'rm </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">toremove</span><span class="p">))</span>
-
- <span class="k">return</span></div>
-</pre></div>
-
- </div>
-
- </div>
- <footer>
-
-
- <hr/>
-
- <div role="contentinfo">
- <p>
-
- © Copyright 2020, Daniel Frisinghelli
-
- </p>
- </div>
-
-
-
- Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a
-
- <a href="https://github.com/rtfd/sphinx_rtd_theme">theme</a>
-
- provided by <a href="https://readthedocs.org">Read the Docs</a>.
-
-</footer>
-
- </div>
- </div>
-
- </section>
-
- </div>
-
-
- <script type="text/javascript">
- jQuery(function () {
- SphinxRtdTheme.Navigation.enable(true);
- });
- </script>
-
-
-
-
-
-
-</body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/_modules/index.html b/docs/_build/html/_modules/index.html
index a46b727b17c6ad03036a0edd32edc2729140b51c..51b13136c6474d895db8c6eea1d85c136451a030 100644
--- a/docs/_build/html/_modules/index.html
+++ b/docs/_build/html/_modules/index.html
@@ -146,21 +146,10 @@
<div itemprop="articleBody">
<h1>All modules for which code is available</h1>
-<ul><li><a href="core/cli.html">core.cli</a></li>
-<li><a href="core/constants.html">core.constants</a></li>
-<li><a href="core/dataset.html">core.dataset</a></li>
-<li><a href="core/graphics.html">core.graphics</a></li>
-<li><a href="core/layers.html">core.layers</a></li>
-<li><a href="core/logging.html">core.logging</a></li>
-<li><a href="core/models.html">core.models</a></li>
-<li><a href="core/predict.html">core.predict</a></li>
-<li><a href="core/split.html">core.split</a></li>
-<li><a href="core/trainer.html">core.trainer</a></li>
-<li><a href="core/transforms.html">core.transforms</a></li>
-<li><a href="core/utils.html">core.utils</a></li>
-<li><a href="pysegcnn/core/dataset.html">pysegcnn.core.dataset</a></li>
+<ul><li><a href="pysegcnn/core/dataset.html">pysegcnn.core.dataset</a></li>
<li><a href="pysegcnn/core/layers.html">pysegcnn.core.layers</a></li>
<li><a href="pysegcnn/core/models.html">pysegcnn.core.models</a></li>
+<li><a href="pysegcnn/core/transforms.html">pysegcnn.core.transforms</a></li>
</ul>
</div>
diff --git a/docs/_build/html/_modules/pysegcnn/core/dataset.html b/docs/_build/html/_modules/pysegcnn/core/dataset.html
index d0811c9a583bd46201ec2b96ddf70cd111ef3e79..d9f8fd6b93a2811c9253779d30b16990a730c43c 100644
--- a/docs/_build/html/_modules/pysegcnn/core/dataset.html
+++ b/docs/_build/html/_modules/pysegcnn/core/dataset.html
@@ -195,54 +195,54 @@
<div class="viewcode-block" id="ImageDataset"><a class="viewcode-back" href="../../../source/generated/pysegcnn.core.dataset.ImageDataset.html#pysegcnn.core.dataset.ImageDataset">[docs]</a><span class="k">class</span> <span class="nc">ImageDataset</span><span class="p">(</span><span class="n">Dataset</span><span class="p">):</span>
- <span class="sa">r</span><span class="sd">"""Base class for multispectral image data.</span>
-
-<span class="sd"> Inheriting from `torch.utils.data.Dataset` to be compliant to the PyTorch</span>
-<span class="sd"> standard. Furthermore, using instances of `torch.utils.data.Dataset`</span>
-<span class="sd"> enables the use of the handy `torch.utils.data.DataLoader` class during</span>
-<span class="sd"> model training.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> root_dir : `str`</span>
-<span class="sd"> The root directory, path to the dataset.</span>
-<span class="sd"> use_bands : `list` [`str`], optional</span>
-<span class="sd"> A list of the spectral bands to use. The default is [].</span>
-<span class="sd"> tile_size : `int` or `None`, optional</span>
-<span class="sd"> The size of the tiles. If not `None`, each scene is divided into square</span>
-<span class="sd"> tiles of shape (tile_size, tile_size). The default is None.</span>
-<span class="sd"> pad : `bool`, optional</span>
-<span class="sd"> Whether to center pad the input image. Set ``pad`` = True, if the</span>
-<span class="sd"> images are not evenly divisible by the ``tile_size``. The image data is</span>
-<span class="sd"> padded with a constant padding value of zero. For each image, the</span>
-<span class="sd"> corresponding ground truth image is padded with a "no data" label.</span>
-<span class="sd"> The default is False.</span>
-<span class="sd"> gt_pattern : `str`, optional</span>
-<span class="sd"> A regural expression to match the ground truth naming convention.</span>
-<span class="sd"> All directories and subdirectories in ``root_dir`` are searched for</span>
-<span class="sd"> files matching ``gt_pattern``. The default is '(.*)gt\\.tif'.</span>
-<span class="sd"> sort : `bool`, optional</span>
-<span class="sd"> Whether to chronologically sort the samples. Useful for time series</span>
-<span class="sd"> data. The default is False.</span>
-<span class="sd"> seed : `int`, optional</span>
-<span class="sd"> The random seed. Used to split the dataset into training, validation</span>
-<span class="sd"> and test set. Useful for reproducibility. The default is 0.</span>
-<span class="sd"> transforms : `list` [`pysegcnn.core.split.Augment`], optional</span>
-<span class="sd"> List of `pysegcnn.core.split.Augment` instances. Each item in</span>
-<span class="sd"> ``transforms`` generates a distinct transformed version of the dataset.</span>
-<span class="sd"> The total dataset is composed of the original untransformed dataset</span>
-<span class="sd"> together with each transformed version of it.</span>
-<span class="sd"> If ``transforms`` = [], only the original dataset is used.</span>
-<span class="sd"> The default is [].</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
+ <span class="sd">"""Base class for multispectral image data.</span>
+
+<span class="sd"> Inheriting from :py:class:`torch.utils.data.Dataset` to be compliant to the</span>
+<span class="sd"> PyTorch standard. This enables the use of the handy</span>
+<span class="sd"> :py:class:`torch.utils.data.DataLoader` class during model training.</span>
<span class="sd"> """</span>
<div class="viewcode-block" id="ImageDataset.__init__"><a class="viewcode-back" href="../../../source/generated/pysegcnn.core.dataset.ImageDataset.html#pysegcnn.core.dataset.ImageDataset.__init__">[docs]</a> <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">root_dir</span><span class="p">,</span> <span class="n">use_bands</span><span class="o">=</span><span class="p">[],</span> <span class="n">tile_size</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">pad</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span>
<span class="n">gt_pattern</span><span class="o">=</span><span class="s1">'(.*)gt.tif'</span><span class="p">,</span> <span class="n">sort</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">0</span><span class="p">,</span> <span class="n">transforms</span><span class="o">=</span><span class="p">[]):</span>
+ <span class="sa">r</span><span class="sd">"""Initialize.</span>
+
+<span class="sd"> Parameters</span>
+<span class="sd"> ----------</span>
+<span class="sd"> root_dir : `str`</span>
+<span class="sd"> The root directory, path to the dataset.</span>
+<span class="sd"> use_bands : `list` [`str`], optional</span>
+<span class="sd"> A list of the spectral bands to use. The default is `[]`.</span>
+<span class="sd"> tile_size : `int` or `None`, optional</span>
+<span class="sd"> The size of the tiles. If not `None`, each scene is divided into</span>
+<span class="sd"> square tiles of shape ``(tile_size, tile_size)``. The default is</span>
+<span class="sd"> `None`.</span>
+<span class="sd"> pad : `bool`, optional</span>
+<span class="sd"> Whether to center pad the input image. Set ``pad=True``, if the</span>
+<span class="sd"> images are not evenly divisible by the ``tile_size``. The image</span>
+<span class="sd"> data is padded with a constant padding value of zero. For each</span>
+<span class="sd"> image, the corresponding ground truth image is padded with a</span>
+<span class="sd"> "no data" label. The default is `False`.</span>
+<span class="sd"> gt_pattern : `str`, optional</span>
+<span class="sd"> A regural expression to match the ground truth naming convention.</span>
+<span class="sd"> All directories and subdirectories in ``root_dir`` are searched for</span>
+<span class="sd"> files matching ``gt_pattern``. The default is `(.*)gt\\.tif`.</span>
+<span class="sd"> sort : `bool`, optional</span>
+<span class="sd"> Whether to chronologically sort the samples. Useful for time series</span>
+<span class="sd"> data. The default is `False`.</span>
+<span class="sd"> seed : `int`, optional</span>
+<span class="sd"> The random seed. Used to split the dataset into training,</span>
+<span class="sd"> validation and test set. Useful for reproducibility. The default is</span>
+<span class="sd"> `0`.</span>
+<span class="sd"> transforms : `list`, optional</span>
+<span class="sd"> List of :py:class:`pysegcnn.core.transforms.Augment` instances.</span>
+<span class="sd"> Each item in ``transforms`` generates a distinct transformed</span>
+<span class="sd"> version of the dataset. The total dataset is composed of the</span>
+<span class="sd"> original untransformed dataset together with each transformed</span>
+<span class="sd"> version of it. If ``transforms=[]``, only the original dataset is</span>
+<span class="sd"> used. The default is `[]`.</span>
+
+<span class="sd"> """</span>
<span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">()</span>
<span class="c1"># dataset configuration</span>
@@ -642,7 +642,7 @@
<span class="sd"> -------</span>
<span class="sd"> stack : `numpy.ndarray`</span>
<span class="sd"> The input data of the sample.</span>
-<span class="sd"> gt : TYPE</span>
+<span class="sd"> gt : `numpy.ndarray`</span>
<span class="sd"> The ground truth of the sample.</span>
<span class="sd"> """</span>
@@ -653,7 +653,7 @@
<span class="k">return</span> <span class="n">stack</span><span class="p">,</span> <span class="n">gt</span>
<span class="k">def</span> <span class="nf">to_tensor</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">x</span><span class="p">,</span> <span class="n">dtype</span><span class="p">):</span>
- <span class="sd">"""Convert ``x`` to `torch.Tensor`.</span>
+ <span class="sd">"""Convert ``x`` to :py:class:`torch.Tensor`.</span>
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
@@ -712,86 +712,43 @@
<span class="k">return</span> <span class="n">fs</span></div>
-<span class="k">class</span> <span class="nc">StandardEoDataset</span><span class="p">(</span><span class="n">ImageDataset</span><span class="p">):</span>
+<div class="viewcode-block" id="StandardEoDataset"><a class="viewcode-back" href="../../../source/generated/pysegcnn.core.dataset.StandardEoDataset.html#pysegcnn.core.dataset.StandardEoDataset">[docs]</a><span class="k">class</span> <span class="nc">StandardEoDataset</span><span class="p">(</span><span class="n">ImageDataset</span><span class="p">):</span>
<span class="sa">r</span><span class="sd">"""Base class for standard Earth Observation style datasets.</span>
-<span class="sd"> `pysegcnn.core.dataset.StandardEoDataset` implements the</span>
-<span class="sd"> `~pysegcnn.core.dataset.StandardEoDataset.compose_scenes` method for</span>
-<span class="sd"> datasets with the following directory structure:</span>
+<span class="sd"> :py:class:`pysegcnn.core.dataset.StandardEoDataset` implements the</span>
+<span class="sd"> :py:meth:`~pysegcnn.core.dataset.StandardEoDataset.compose_scenes` method</span>
+<span class="sd"> for datasets with the following directory structure:</span>
<span class="sd"> root_dir/</span>
-<span class="sd"> scene_id_1/</span>
-<span class="sd"> scene_id_1_B1.tif</span>
-<span class="sd"> scene_id_1_B2.tif</span>
-<span class="sd"> .</span>
-<span class="sd"> .</span>
-<span class="sd"> .</span>
-<span class="sd"> scene_id_1_BN.tif</span>
-<span class="sd"> scene_id_2/</span>
-<span class="sd"> scene_id_2_B1.tif</span>
-<span class="sd"> scene_id_2_B2.tif</span>
-<span class="sd"> .</span>
-<span class="sd"> .</span>
-<span class="sd"> .</span>
-<span class="sd"> scene_id_2_BN.tif</span>
-<span class="sd"> .</span>
-<span class="sd"> .</span>
-<span class="sd"> .</span>
-<span class="sd"> scene_id_N/</span>
-<span class="sd"> .</span>
-<span class="sd"> .</span>
-<span class="sd"> .</span>
+<span class="sd"> - scene_id_1/</span>
+<span class="sd"> - scene_id_1_B1.tif</span>
+<span class="sd"> - scene_id_1_B2.tif</span>
+<span class="sd"> - ...</span>
+<span class="sd"> - scene_id_1_BN.tif</span>
+<span class="sd"> - scene_id_2/</span>
+<span class="sd"> - scene_id_2_B1.tif</span>
+<span class="sd"> - scene_id_2_B2.tif</span>
+<span class="sd"> - ...</span>
+<span class="sd"> - scene_id_2_BN.tif</span>
+<span class="sd"> - ...</span>
+<span class="sd"> - scene_id_N/</span>
+<span class="sd"> - ...</span>
<span class="sd"> If your dataset shares this directory structure, you can directly inherit</span>
-<span class="sd"> `pysegcnn.core.dataset.StandardEoDataset` and implement the remaining</span>
-<span class="sd"> methods.</span>
-
-<span class="sd"> See `pysegcnn.core.dataset.SparcsDataset` for an example.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> root_dir : `str`</span>
-<span class="sd"> The root directory, path to the dataset.</span>
-<span class="sd"> use_bands : `list` [`str`], optional</span>
-<span class="sd"> A list of the spectral bands to use. The default is [].</span>
-<span class="sd"> tile_size : `int` or `None`, optional</span>
-<span class="sd"> The size of the tiles. If not `None`, each scene is divided into square</span>
-<span class="sd"> tiles of shape (tile_size, tile_size). The default is None.</span>
-<span class="sd"> pad : `bool`, optional</span>
-<span class="sd"> Whether to center pad the input image. Set ``pad`` = True, if the</span>
-<span class="sd"> images are not evenly divisible by the ``tile_size``. The image data is</span>
-<span class="sd"> padded with a constant padding value of zero. For each image, the</span>
-<span class="sd"> corresponding ground truth image is padded with a "no data" label.</span>
-<span class="sd"> The default is False.</span>
-<span class="sd"> gt_pattern : `str`, optional</span>
-<span class="sd"> A regural expression to match the ground truth naming convention. All</span>
-<span class="sd"> directories and subdirectories in ``root_dir`` are searched for files</span>
-<span class="sd"> matching ``gt_pattern``. The default is '(.*)gt\\.tif'.</span>
-<span class="sd"> sort : `bool`, optional</span>
-<span class="sd"> Whether to chronologically sort the samples. Useful for time series</span>
-<span class="sd"> data. The default is False.</span>
-<span class="sd"> seed : `int`, optional</span>
-<span class="sd"> The random seed. Used to split the dataset into training, validation</span>
-<span class="sd"> and test set. Useful for reproducibility. The default is 0.</span>
-<span class="sd"> transforms : `list` [`pysegcnn.core.split.Augment`], optional</span>
-<span class="sd"> List of `pysegcnn.core.split.Augment` instances. Each item in</span>
-<span class="sd"> ``transforms`` generates a distinct transformed version of the dataset.</span>
-<span class="sd"> The total dataset is composed of the original untransformed dataset</span>
-<span class="sd"> together with each transformed version of it.</span>
-<span class="sd"> If ``transforms`` = [], only the original dataset is used.</span>
-<span class="sd"> The default is [].</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
+<span class="sd"> :py:class:`pysegcnn.core.dataset.StandardEoDataset` and implement the</span>
+<span class="sd"> remaining methods. If not, you can use</span>
+<span class="sd"> :py:func:`pysegcnn.core.utils.standard_eo_structure` to transfer your</span>
+<span class="sd"> dataset to the above directory structure.</span>
+
+<span class="sd"> See :py:class:`pysegcnn.core.dataset.SparcsDataset` for an example.</span>
<span class="sd"> """</span>
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">root_dir</span><span class="p">,</span> <span class="n">use_bands</span><span class="o">=</span><span class="p">[],</span> <span class="n">tile_size</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">pad</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span>
+<div class="viewcode-block" id="StandardEoDataset.__init__"><a class="viewcode-back" href="../../../source/generated/pysegcnn.core.dataset.StandardEoDataset.html#pysegcnn.core.dataset.StandardEoDataset.__init__">[docs]</a> <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">root_dir</span><span class="p">,</span> <span class="n">use_bands</span><span class="o">=</span><span class="p">[],</span> <span class="n">tile_size</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">pad</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span>
<span class="n">gt_pattern</span><span class="o">=</span><span class="s1">'(.*)gt</span><span class="se">\\</span><span class="s1">.tif'</span><span class="p">,</span> <span class="n">sort</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">0</span><span class="p">,</span> <span class="n">transforms</span><span class="o">=</span><span class="p">[]):</span>
<span class="c1"># initialize super class ImageDataset</span>
<span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">root_dir</span><span class="p">,</span> <span class="n">use_bands</span><span class="p">,</span> <span class="n">tile_size</span><span class="p">,</span> <span class="n">pad</span><span class="p">,</span> <span class="n">gt_pattern</span><span class="p">,</span>
- <span class="n">sort</span><span class="p">,</span> <span class="n">seed</span><span class="p">,</span> <span class="n">transforms</span><span class="p">)</span>
+ <span class="n">sort</span><span class="p">,</span> <span class="n">seed</span><span class="p">,</span> <span class="n">transforms</span><span class="p">)</span></div>
<span class="k">def</span> <span class="nf">_get_band_number</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">path</span><span class="p">):</span>
<span class="sd">"""Return the band number of a scene .tif file.</span>
@@ -859,32 +816,7 @@
<span class="k">return</span> <span class="n">scene_data</span>
<span class="k">def</span> <span class="nf">compose_scenes</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Build the list of samples of the dataset.</span>
-
-<span class="sd"> Each sample is represented by a dictionary.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> scenes : `list` [`dict`]</span>
-<span class="sd"> Each item in ``scenes`` is a `dict` with keys:</span>
-<span class="sd"> ``'band_name_1'``</span>
-<span class="sd"> Path to the file of band_1.</span>
-<span class="sd"> ``'band_name_2'``</span>
-<span class="sd"> Path to the file of band_2.</span>
-<span class="sd"> ``'band_name_n'``</span>
-<span class="sd"> Path to the file of band_n.</span>
-<span class="sd"> ``'gt'``</span>
-<span class="sd"> Path to the ground truth file.</span>
-<span class="sd"> ``'date'``</span>
-<span class="sd"> The date of the sample.</span>
-<span class="sd"> ``'tile'``</span>
-<span class="sd"> The tile id of the sample.</span>
-<span class="sd"> ``'transform'``</span>
-<span class="sd"> The transformation to apply.</span>
-<span class="sd"> ``'id'``</span>
-<span class="sd"> The scene identifier.</span>
-
-<span class="sd"> """</span>
+ <span class="sd">"""Build the list of samples of the dataset."""</span>
<span class="c1"># search the root directory</span>
<span class="n">scenes</span> <span class="o">=</span> <span class="p">[]</span>
<span class="bp">self</span><span class="o">.</span><span class="n">gt</span> <span class="o">=</span> <span class="p">[]</span>
@@ -904,7 +836,7 @@
<span class="c1"># list the spectral bands of the scene</span>
<span class="n">bands</span> <span class="o">=</span> <span class="p">[</span><span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">dirpath</span><span class="p">,</span> <span class="n">f</span><span class="p">)</span> <span class="k">for</span> <span class="n">f</span> <span class="ow">in</span> <span class="n">files</span>
- <span class="k">if</span> <span class="n">re</span><span class="o">.</span><span class="n">search</span><span class="p">(</span><span class="s1">'B</span><span class="se">\\</span><span class="s1">dA|B</span><span class="se">\\</span><span class="s1">d{1,2}'</span><span class="p">,</span> <span class="n">f</span><span class="p">)]</span>
+ <span class="k">if</span> <span class="n">re</span><span class="o">.</span><span class="n">search</span><span class="p">(</span><span class="s1">'B</span><span class="se">\\</span><span class="s1">dA|B</span><span class="se">\\</span><span class="s1">d{1,2}(.*).tif$'</span><span class="p">,</span> <span class="n">f</span><span class="p">)]</span>
<span class="c1"># get the ground truth mask</span>
<span class="k">try</span><span class="p">:</span>
@@ -946,51 +878,17 @@
<span class="n">scenes</span><span class="o">.</span><span class="n">sort</span><span class="p">(</span><span class="n">key</span><span class="o">=</span><span class="k">lambda</span> <span class="n">k</span><span class="p">:</span> <span class="n">k</span><span class="p">[</span><span class="s1">'date'</span><span class="p">])</span>
<span class="bp">self</span><span class="o">.</span><span class="n">gt</span><span class="o">.</span><span class="n">sort</span><span class="p">(</span><span class="n">key</span><span class="o">=</span><span class="k">lambda</span> <span class="n">k</span><span class="p">:</span> <span class="bp">self</span><span class="o">.</span><span class="n">parse_scene_id</span><span class="p">(</span><span class="n">k</span><span class="p">)[</span><span class="s1">'date'</span><span class="p">])</span>
- <span class="k">return</span> <span class="n">scenes</span>
+ <span class="k">return</span> <span class="n">scenes</span></div>
<div class="viewcode-block" id="SparcsDataset"><a class="viewcode-back" href="../../../source/generated/pysegcnn.core.dataset.SparcsDataset.html#pysegcnn.core.dataset.SparcsDataset">[docs]</a><span class="k">class</span> <span class="nc">SparcsDataset</span><span class="p">(</span><span class="n">StandardEoDataset</span><span class="p">):</span>
- <span class="sa">r</span><span class="sd">"""Class for the `Sparcs`_ dataset.</span>
+ <span class="sd">"""Class for the `Sparcs`_ dataset by `Hughes & Hayes (2014)`_.</span>
<span class="sd"> .. _Sparcs:</span>
<span class="sd"> https://www.usgs.gov/land-resources/nli/landsat/spatial-procedures-automated-removal-cloud-and-shadow-sparcs-validation</span>
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> root_dir : `str`</span>
-<span class="sd"> The root directory, path to the dataset.</span>
-<span class="sd"> use_bands : `list` [`str`], optional</span>
-<span class="sd"> A list of the spectral bands to use. The default is [].</span>
-<span class="sd"> tile_size : `int` or `None`, optional</span>
-<span class="sd"> The size of the tiles. If not `None`, each scene is divided into square</span>
-<span class="sd"> tiles of shape (tile_size, tile_size). The default is None.</span>
-<span class="sd"> pad : `bool`, optional</span>
-<span class="sd"> Whether to center pad the input image. Set ``pad`` = True, if the</span>
-<span class="sd"> images are not evenly divisible by the ``tile_size``. The image data is</span>
-<span class="sd"> padded with a constant padding value of zero. For each image, the</span>
-<span class="sd"> corresponding ground truth image is padded with a "no data" label.</span>
-<span class="sd"> The default is False.</span>
-<span class="sd"> gt_pattern : `str`, optional</span>
-<span class="sd"> A regural expression to match the ground truth naming convention. All</span>
-<span class="sd"> directories and subdirectories in ``root_dir`` are searched for files</span>
-<span class="sd"> matching ``gt_pattern``. The default is '(.*)gt\\.tif'.</span>
-<span class="sd"> sort : `bool`, optional</span>
-<span class="sd"> Whether to chronologically sort the samples. Useful for time series</span>
-<span class="sd"> data. The default is False.</span>
-<span class="sd"> seed : `int`, optional</span>
-<span class="sd"> The random seed. Used to split the dataset into training, validation</span>
-<span class="sd"> and test set. Useful for reproducibility. The default is 0.</span>
-<span class="sd"> transforms : `list` [`pysegcnn.core.split.Augment`], optional</span>
-<span class="sd"> List of `pysegcnn.core.split.Augment` instances. Each item in</span>
-<span class="sd"> ``transforms`` generates a distinct transformed version of the dataset.</span>
-<span class="sd"> The total dataset is composed of the original untransformed dataset</span>
-<span class="sd"> together with each transformed version of it.</span>
-<span class="sd"> If ``transforms`` = [], only the original dataset is used.</span>
-<span class="sd"> The default is [].</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
+<span class="sd"> .. _Hughes & Hayes (2014):</span>
+<span class="sd"> https://www.mdpi.com/2072-4292/6/6/4907</span>
<span class="sd"> """</span>
@@ -1073,46 +971,7 @@
<span class="k">class</span> <span class="nc">ProSnowDataset</span><span class="p">(</span><span class="n">StandardEoDataset</span><span class="p">):</span>
- <span class="sa">r</span><span class="sd">"""Class for the ProSnow datasets.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> root_dir : `str`</span>
-<span class="sd"> The root directory, path to the dataset.</span>
-<span class="sd"> use_bands : `list` [`str`], optional</span>
-<span class="sd"> A list of the spectral bands to use. The default is [].</span>
-<span class="sd"> tile_size : `int` or `None`, optional</span>
-<span class="sd"> The size of the tiles. If not `None`, each scene is divided into square</span>
-<span class="sd"> tiles of shape (tile_size, tile_size). The default is None.</span>
-<span class="sd"> pad : `bool`, optional</span>
-<span class="sd"> Whether to center pad the input image. Set ``pad`` = True, if the</span>
-<span class="sd"> images are not evenly divisible by the ``tile_size``. The image data is</span>
-<span class="sd"> padded with a constant padding value of zero. For each image, the</span>
-<span class="sd"> corresponding ground truth image is padded with a "no data" label.</span>
-<span class="sd"> The default is False.</span>
-<span class="sd"> gt_pattern : `str`, optional</span>
-<span class="sd"> A regural expression to match the ground truth naming convention. All</span>
-<span class="sd"> directories and subdirectories in ``root_dir`` are searched for files</span>
-<span class="sd"> matching ``gt_pattern``. The default is '(.*)gt\\.tif'.</span>
-<span class="sd"> sort : `bool`, optional</span>
-<span class="sd"> Whether to chronologically sort the samples. Useful for time series</span>
-<span class="sd"> data. The default is False.</span>
-<span class="sd"> seed : `int`, optional</span>
-<span class="sd"> The random seed. Used to split the dataset into training, validation</span>
-<span class="sd"> and test set. Useful for reproducibility. The default is 0.</span>
-<span class="sd"> transforms : `list` [`pysegcnn.core.split.Augment`], optional</span>
-<span class="sd"> List of `pysegcnn.core.split.Augment` instances. Each item in</span>
-<span class="sd"> ``transforms`` generates a distinct transformed version of the dataset.</span>
-<span class="sd"> The total dataset is composed of the original untransformed dataset</span>
-<span class="sd"> together with each transformed version of it.</span>
-<span class="sd"> If ``transforms`` = [], only the original dataset is used.</span>
-<span class="sd"> The default is [].</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
+ <span class="sd">"""Class for the ProSnow datasets."""</span>
<span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">root_dir</span><span class="p">,</span> <span class="n">use_bands</span><span class="o">=</span><span class="p">[],</span> <span class="n">tile_size</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">pad</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span>
<span class="n">gt_pattern</span><span class="o">=</span><span class="s1">'(.*)gt</span><span class="se">\\</span><span class="s1">.tif'</span><span class="p">,</span> <span class="n">sort</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">0</span><span class="p">,</span> <span class="n">transforms</span><span class="o">=</span><span class="p">[]):</span>
@@ -1182,46 +1041,7 @@
<span class="k">class</span> <span class="nc">ProSnowGarmisch</span><span class="p">(</span><span class="n">ProSnowDataset</span><span class="p">):</span>
- <span class="sa">r</span><span class="sd">"""Class for the ProSnow Garmisch dataset.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> root_dir : `str`</span>
-<span class="sd"> The root directory, path to the dataset.</span>
-<span class="sd"> use_bands : `list` [`str`], optional</span>
-<span class="sd"> A list of the spectral bands to use. The default is [].</span>
-<span class="sd"> tile_size : `int` or `None`, optional</span>
-<span class="sd"> The size of the tiles. If not `None`, each scene is divided into square</span>
-<span class="sd"> tiles of shape (tile_size, tile_size). The default is None.</span>
-<span class="sd"> pad : `bool`, optional</span>
-<span class="sd"> Whether to center pad the input image. Set ``pad`` = True, if the</span>
-<span class="sd"> images are not evenly divisible by the ``tile_size``. The image data is</span>
-<span class="sd"> padded with a constant padding value of zero. For each image, the</span>
-<span class="sd"> corresponding ground truth image is padded with a "no data" label.</span>
-<span class="sd"> The default is False.</span>
-<span class="sd"> gt_pattern : `str`, optional</span>
-<span class="sd"> A regural expression to match the ground truth naming convention. All</span>
-<span class="sd"> directories and subdirectories in ``root_dir`` are searched for files</span>
-<span class="sd"> matching ``gt_pattern``. The default is '(.*)gt\\.tif'.</span>
-<span class="sd"> sort : `bool`, optional</span>
-<span class="sd"> Whether to chronologically sort the samples. Useful for time series</span>
-<span class="sd"> data. The default is False.</span>
-<span class="sd"> seed : `int`, optional</span>
-<span class="sd"> The random seed. Used to split the dataset into training, validation</span>
-<span class="sd"> and test set. Useful for reproducibility. The default is 0.</span>
-<span class="sd"> transforms : `list` [`pysegcnn.core.split.Augment`], optional</span>
-<span class="sd"> List of `pysegcnn.core.split.Augment` instances. Each item in</span>
-<span class="sd"> ``transforms`` generates a distinct transformed version of the dataset.</span>
-<span class="sd"> The total dataset is composed of the original untransformed dataset</span>
-<span class="sd"> together with each transformed version of it.</span>
-<span class="sd"> If ``transforms`` = [], only the original dataset is used.</span>
-<span class="sd"> The default is [].</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
+ <span class="sd">"""Class for the ProSnow Garmisch dataset."""</span>
<span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">root_dir</span><span class="p">,</span> <span class="n">use_bands</span><span class="o">=</span><span class="p">[],</span> <span class="n">tile_size</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">pad</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span>
<span class="n">gt_pattern</span><span class="o">=</span><span class="s1">'(.*)gt</span><span class="se">\\</span><span class="s1">.tif'</span><span class="p">,</span> <span class="n">sort</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">0</span><span class="p">,</span> <span class="n">transforms</span><span class="o">=</span><span class="p">[]):</span>
@@ -1242,46 +1062,7 @@
<span class="k">class</span> <span class="nc">ProSnowObergurgl</span><span class="p">(</span><span class="n">ProSnowDataset</span><span class="p">):</span>
- <span class="sa">r</span><span class="sd">"""Class for the ProSnow Obergurgl dataset.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> root_dir : `str`</span>
-<span class="sd"> The root directory, path to the dataset.</span>
-<span class="sd"> use_bands : `list` [`str`], optional</span>
-<span class="sd"> A list of the spectral bands to use. The default is [].</span>
-<span class="sd"> tile_size : `int` or `None`, optional</span>
-<span class="sd"> The size of the tiles. If not `None`, each scene is divided into square</span>
-<span class="sd"> tiles of shape (tile_size, tile_size). The default is None.</span>
-<span class="sd"> pad : `bool`, optional</span>
-<span class="sd"> Whether to center pad the input image. Set ``pad`` = True, if the</span>
-<span class="sd"> images are not evenly divisible by the ``tile_size``. The image data is</span>
-<span class="sd"> padded with a constant padding value of zero. For each image, the</span>
-<span class="sd"> corresponding ground truth image is padded with a "no data" label.</span>
-<span class="sd"> The default is False.</span>
-<span class="sd"> gt_pattern : `str`, optional</span>
-<span class="sd"> A regural expression to match the ground truth naming convention. All</span>
-<span class="sd"> directories and subdirectories in ``root_dir`` are searched for files</span>
-<span class="sd"> matching ``gt_pattern``. The default is '(.*)gt\\.tif'.</span>
-<span class="sd"> sort : `bool`, optional</span>
-<span class="sd"> Whether to chronologically sort the samples. Useful for time series</span>
-<span class="sd"> data. The default is False.</span>
-<span class="sd"> seed : `int`, optional</span>
-<span class="sd"> The random seed. Used to split the dataset into training, validation</span>
-<span class="sd"> and test set. Useful for reproducibility. The default is 0.</span>
-<span class="sd"> transforms : `list` [`pysegcnn.core.split.Augment`], optional</span>
-<span class="sd"> List of `pysegcnn.core.split.Augment` instances. Each item in</span>
-<span class="sd"> ``transforms`` generates a distinct transformed version of the dataset.</span>
-<span class="sd"> The total dataset is composed of the original untransformed dataset</span>
-<span class="sd"> together with each transformed version of it.</span>
-<span class="sd"> If ``transforms`` = [], only the original dataset is used.</span>
-<span class="sd"> The default is [].</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
+ <span class="sd">"""Class for the ProSnow Obergurgl dataset."""</span>
<span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">root_dir</span><span class="p">,</span> <span class="n">use_bands</span><span class="o">=</span><span class="p">[],</span> <span class="n">tile_size</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">pad</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span>
<span class="n">gt_pattern</span><span class="o">=</span><span class="s1">'(.*)gt</span><span class="se">\\</span><span class="s1">.tif'</span><span class="p">,</span> <span class="n">sort</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">0</span><span class="p">,</span> <span class="n">transforms</span><span class="o">=</span><span class="p">[]):</span>
@@ -1302,50 +1083,13 @@
<div class="viewcode-block" id="Cloud95Dataset"><a class="viewcode-back" href="../../../source/generated/pysegcnn.core.dataset.Cloud95Dataset.html#pysegcnn.core.dataset.Cloud95Dataset">[docs]</a><span class="k">class</span> <span class="nc">Cloud95Dataset</span><span class="p">(</span><span class="n">ImageDataset</span><span class="p">):</span>
- <span class="sa">r</span><span class="sd">"""Class for the `Cloud-95`_ dataset by `Mohajerani & Saeedi (2020)`_.</span>
+ <span class="sd">"""Class for the `Cloud-95`_ dataset by `Mohajerani & Saeedi (2020)`_.</span>
<span class="sd"> .. _Cloud-95:</span>
<span class="sd"> https://github.com/SorourMo/95-Cloud-An-Extension-to-38-Cloud-Dataset</span>
<span class="sd"> .. _Mohajerani & Saeedi (2020):</span>
<span class="sd"> https://arxiv.org/abs/2001.08768</span>
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> root_dir : `str`</span>
-<span class="sd"> The root directory, path to the dataset.</span>
-<span class="sd"> use_bands : `list` [`str`], optional</span>
-<span class="sd"> A list of the spectral bands to use. The default is [].</span>
-<span class="sd"> tile_size : `int` or `None`, optional</span>
-<span class="sd"> The size of the tiles. If not `None`, each scene is divided into square</span>
-<span class="sd"> tiles of shape (tile_size, tile_size). The default is None.</span>
-<span class="sd"> pad : `bool`, optional</span>
-<span class="sd"> Whether to center pad the input image. Set ``pad`` = True, if the</span>
-<span class="sd"> images are not evenly divisible by the ``tile_size``. The image data is</span>
-<span class="sd"> padded with a constant padding value of zero. For each image, the</span>
-<span class="sd"> corresponding ground truth image is padded with a "no data" label.</span>
-<span class="sd"> The default is False.</span>
-<span class="sd"> gt_pattern : `str`, optional</span>
-<span class="sd"> A regural expression to match the ground truth naming convention. All</span>
-<span class="sd"> directories and subdirectories in ``root_dir`` are searched for files</span>
-<span class="sd"> matching ``gt_pattern``. The default is '(.*)gt\\.tif'.</span>
-<span class="sd"> sort : `bool`, optional</span>
-<span class="sd"> Whether to chronologically sort the samples. Useful for time series</span>
-<span class="sd"> data. The default is False.</span>
-<span class="sd"> seed : `int`, optional</span>
-<span class="sd"> The random seed. Used to split the dataset into training, validation</span>
-<span class="sd"> and test set. Useful for reproducibility. The default is 0.</span>
-<span class="sd"> transforms : `list` [`pysegcnn.core.split.Augment`], optional</span>
-<span class="sd"> List of `pysegcnn.core.split.Augment` instances. Each item in</span>
-<span class="sd"> ``transforms`` generates a distinct transformed version of the dataset.</span>
-<span class="sd"> The total dataset is composed of the original untransformed dataset</span>
-<span class="sd"> together with each transformed version of it.</span>
-<span class="sd"> If ``transforms`` = [], only the original dataset is used.</span>
-<span class="sd"> The default is [].</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
<span class="sd"> """</span>
<div class="viewcode-block" id="Cloud95Dataset.__init__"><a class="viewcode-back" href="../../../source/generated/pysegcnn.core.dataset.Cloud95Dataset.html#pysegcnn.core.dataset.Cloud95Dataset.__init__">[docs]</a> <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">root_dir</span><span class="p">,</span> <span class="n">use_bands</span><span class="o">=</span><span class="p">[],</span> <span class="n">tile_size</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">pad</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span>
@@ -1435,32 +1179,7 @@
<span class="k">return</span> <span class="n">parse_landsat_scene</span><span class="p">(</span><span class="n">scene_id</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">compose_scenes</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""Build the list of samples of the dataset.</span>
-
-<span class="sd"> Each sample is represented by a dictionary.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> scenes : `list` [`dict`]</span>
-<span class="sd"> Each item in ``scenes`` is a `dict` with keys:</span>
-<span class="sd"> ``'band_name_1'``</span>
-<span class="sd"> Path to the file of band_1.</span>
-<span class="sd"> ``'band_name_2'``</span>
-<span class="sd"> Path to the file of band_2.</span>
-<span class="sd"> ``'band_name_n'``</span>
-<span class="sd"> Path to the file of band_n.</span>
-<span class="sd"> ``'gt'``</span>
-<span class="sd"> Path to the ground truth file.</span>
-<span class="sd"> ``'date'``</span>
-<span class="sd"> The date of the sample.</span>
-<span class="sd"> ``'tile'``</span>
-<span class="sd"> The tile id of the sample.</span>
-<span class="sd"> ``'transform'``</span>
-<span class="sd"> The transformation to apply.</span>
-<span class="sd"> ``'id'``</span>
-<span class="sd"> The scene identifier.</span>
-
-<span class="sd"> """</span>
+ <span class="sd">"""Build the list of samples of the dataset."""</span>
<span class="c1"># whether to exclude patches with more than 80% black pixels</span>
<span class="n">ipatches</span> <span class="o">=</span> <span class="p">[]</span>
<span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">exclude</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
diff --git a/docs/_build/html/_modules/pysegcnn/core/layers.html b/docs/_build/html/_modules/pysegcnn/core/layers.html
index 1c7ad933a346f350b49e1ade27c8a357ec2d6ef7..b0373e234919b4f1c3c6f900a6fa0d2105cef33e 100644
--- a/docs/_build/html/_modules/pysegcnn/core/layers.html
+++ b/docs/_build/html/_modules/pysegcnn/core/layers.html
@@ -178,29 +178,30 @@
<span class="sd"> calculated such that the output of the convolution has the same spatial</span>
<span class="sd"> dimensions as the input.</span>
-<span class="sd"> Parameters</span>
+<span class="sd"> Attributes</span>
<span class="sd"> ----------</span>
-<span class="sd"> *args: `list` [`str`]</span>
-<span class="sd"> positional arguments passed to `torch.nn.Conv2d`:</span>
-<span class="sd"> ``'in_channels'``: `int`</span>
-<span class="sd"> Number of input channels.</span>
-<span class="sd"> ``'out_channels'``: `int`</span>
-<span class="sd"> Number of output channels.</span>
-<span class="sd"> ``'kernel_size'``: `int` or `tuple` [`int`]</span>
-<span class="sd"> Size of the convolving kernel.</span>
-<span class="sd"> **kwargs: 'dict' [`str`]</span>
-<span class="sd"> Additional keyword arguments passed to `torch.nn.Conv2d`_.</span>
-
-<span class="sd"> .. _torch.nn.Conv2d:</span>
-<span class="sd"> https://pytorch.org/docs/stable/generated/torch.nn.Conv2d.html#torch.nn.Conv2d</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
+<span class="sd"> padding : `tuple` [`int`]</span>
+<span class="sd"> The amount of padding, (pad_height, pad_width).</span>
<span class="sd"> """</span>
<div class="viewcode-block" id="Conv2dSame.__init__"><a class="viewcode-back" href="../../../source/generated/pysegcnn.core.layers.Conv2dSame.html#pysegcnn.core.layers.Conv2dSame.__init__">[docs]</a> <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="o">*</span><span class="n">args</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
+ <span class="sd">"""Initialize.</span>
+
+<span class="sd"> Parameters</span>
+<span class="sd"> ----------</span>
+<span class="sd"> *args: `list` [`str`]</span>
+<span class="sd"> positional arguments passed to :py:class:`torch.nn.Conv2d`:</span>
+<span class="sd"> ``'in_channels'``: `int`</span>
+<span class="sd"> Number of input channels.</span>
+<span class="sd"> ``'out_channels'``: `int`</span>
+<span class="sd"> Number of output channels.</span>
+<span class="sd"> ``'kernel_size'``: `int` or `tuple` [`int`]</span>
+<span class="sd"> Size of the convolving kernel.</span>
+<span class="sd"> **kwargs: `dict` [`str`]</span>
+<span class="sd"> Additional keyword arguments passed to :py:class:`torch.nn.Conv2d`.</span>
+
+<span class="sd"> """</span>
<span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="o">*</span><span class="n">args</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span>
<span class="c1"># define tensorflows "SAME" padding for stride = 1</span>
@@ -223,7 +224,7 @@
<span class="sd"> Returns</span>
<span class="sd"> -------</span>
<span class="sd"> p : `int`</span>
-<span class="sd"> the amount of padding.</span>
+<span class="sd"> The amount of padding.</span>
<span class="sd"> """</span>
<span class="c1"># calculates the padding so that the convolution</span>
@@ -240,7 +241,7 @@
<span class="sd"> Number of input channels.</span>
<span class="sd"> out_channels : `int`</span>
<span class="sd"> Number of output channels.</span>
-<span class="sd"> **kwargs: 'dict' [`str`]</span>
+<span class="sd"> **kwargs: `dict` [`str`]</span>
<span class="sd"> Additional arguments passed to `pysegcnn.core.layers.Conv2dSame`.</span>
<span class="sd"> Returns</span>
@@ -263,28 +264,39 @@
<div class="viewcode-block" id="Block"><a class="viewcode-back" href="../../../source/generated/pysegcnn.core.layers.Block.html#pysegcnn.core.layers.Block">[docs]</a><span class="k">class</span> <span class="nc">Block</span><span class="p">(</span><span class="n">nn</span><span class="o">.</span><span class="n">Module</span><span class="p">):</span>
<span class="sd">"""Basic convolutional block.</span>
-<span class="sd"> Parameters</span>
+<span class="sd"> Attributes</span>
<span class="sd"> ----------</span>
<span class="sd"> in_channels : `int`</span>
<span class="sd"> Number of input channels.</span>
<span class="sd"> out_channels : `int`</span>
<span class="sd"> Number of output channels.</span>
-<span class="sd"> **kwargs: 'dict' [`str`]</span>
-<span class="sd"> Additional arguments passed to `pysegcnn.core.layers.Conv2dSame`.</span>
-
-<span class="sd"> Raises</span>
-<span class="sd"> ------</span>
-<span class="sd"> TypeError</span>
-<span class="sd"> Raised if `~pysegcnn.core.layers.Block.layers` method does not return</span>
-<span class="sd"> an instance of `torch.nn.Sequential`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
+<span class="sd"> kwargs: `dict` [`str`]</span>
+<span class="sd"> Additional arguments passed to</span>
+<span class="sd"> :py:class:`pysegcnn.core.layers.Conv2dSame`.</span>
+<span class="sd"> conv : :py:class:`torch.nn.Sequential`</span>
+<span class="sd"> The convolutional layers of the block.</span>
<span class="sd"> """</span>
<div class="viewcode-block" id="Block.__init__"><a class="viewcode-back" href="../../../source/generated/pysegcnn.core.layers.Block.html#pysegcnn.core.layers.Block.__init__">[docs]</a> <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">in_channels</span><span class="p">,</span> <span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
+ <span class="sd">"""Initialize.</span>
+
+<span class="sd"> Parameters</span>
+<span class="sd"> ----------</span>
+<span class="sd"> in_channels : `int`</span>
+<span class="sd"> Number of input channels.</span>
+<span class="sd"> out_channels : `int`</span>
+<span class="sd"> Number of output channels.</span>
+<span class="sd"> **kwargs: `dict` [`str`]</span>
+<span class="sd"> Additional arguments passed to</span>
+<span class="sd"> :py:class:`pysegcnn.core.layers.Conv2dSame`.</span>
+
+<span class="sd"> Raises</span>
+<span class="sd"> ------</span>
+<span class="sd"> TypeError</span>
+<span class="sd"> Raised if :py:meth:`~pysegcnn.core.layers.Block.layers` method does</span>
+<span class="sd"> not return an instance of :py:class:`torch.nn.Sequential`.</span>
+<span class="sd"> """</span>
<span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">()</span>
<span class="c1"># number of input and output channels</span>
@@ -307,13 +319,13 @@
<span class="sd"> Raises</span>
<span class="sd"> ------</span>
<span class="sd"> NotImplementedError</span>
-<span class="sd"> Raised if `pysegcnn.core.layers.Block` is not inherited.</span>
+<span class="sd"> Raised if :py:class:`pysegcnn.core.layers.Block` is not inherited.</span>
<span class="sd"> Returns</span>
<span class="sd"> -------</span>
-<span class="sd"> layers : `torch.nn.Sequential` [`torch.nn.Module`]</span>
-<span class="sd"> Return an instance of `torch.nn.Sequential` containing a sequence</span>
-<span class="sd"> of layer (`torch.nn.Module` ) instances.</span>
+<span class="sd"> layers : :py:class:`torch.nn.Sequential` [:py:class:`torch.nn.Module`]</span>
+<span class="sd"> Return an instance of :py:class:`torch.nn.Sequential` containing a</span>
+<span class="sd"> sequence of layer (:py:class:`torch.nn.Module` ) instances.</span>
<span class="sd"> """</span>
<span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">(</span><span class="s1">'Return an instance of </span><span class="si">{}</span><span class="s1">.'</span>
@@ -325,33 +337,14 @@
<span class="sd"> Raises</span>
<span class="sd"> ------</span>
<span class="sd"> NotImplementedError</span>
-<span class="sd"> Raised if `pysegcnn.core.layers.Block` is not inherited.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
+<span class="sd"> Raised if :py:class:`pysegcnn.core.layers.Block` is not inherited.</span>
<span class="sd"> """</span>
<span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">(</span><span class="s1">'Implement the forward pass.'</span><span class="p">)</span></div>
<div class="viewcode-block" id="EncoderBlock"><a class="viewcode-back" href="../../../source/generated/pysegcnn.core.layers.EncoderBlock.html#pysegcnn.core.layers.EncoderBlock">[docs]</a><span class="k">class</span> <span class="nc">EncoderBlock</span><span class="p">(</span><span class="n">Block</span><span class="p">):</span>
- <span class="sd">"""Block of a convolutional encoder.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> in_channels : `int`</span>
-<span class="sd"> Number of input channels.</span>
-<span class="sd"> out_channels : `int`</span>
-<span class="sd"> Number of output channels.</span>
-<span class="sd"> **kwargs: 'dict' [`str`]</span>
-<span class="sd"> Additional arguments passed to `pysegcnn.core.layers.Conv2dSame`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
+ <span class="sd">"""Block of a convolutional encoder."""</span>
<div class="viewcode-block" id="EncoderBlock.__init__"><a class="viewcode-back" href="../../../source/generated/pysegcnn.core.layers.EncoderBlock.html#pysegcnn.core.layers.EncoderBlock.__init__">[docs]</a> <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">in_channels</span><span class="p">,</span> <span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
<span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">in_channels</span><span class="p">,</span> <span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span></div>
@@ -361,21 +354,21 @@
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
-<span class="sd"> x : `torch.Tensor`</span>
+<span class="sd"> x : :py:class:`torch.Tensor`</span>
<span class="sd"> Input tensor, e.g. output of the previous block/layer.</span>
<span class="sd"> Returns</span>
<span class="sd"> -------</span>
-<span class="sd"> y : `torch.Tensor`, shape=(batch, channel, height, width)</span>
+<span class="sd"> y : :py:class:`torch.Tensor`, shape=(b, c, h, w)</span>
<span class="sd"> Output of the encoder block.</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
+<span class="sd"> x : :py:class:`torch.Tensor`, shape=(b, c, h, w)</span>
<span class="sd"> Intermediate output before applying downsampling. Useful to</span>
<span class="sd"> implement skip connections.</span>
-<span class="sd"> indices : `torch.Tensor` or `None`</span>
+<span class="sd"> indices : :py:class:`torch.Tensor` or `None`</span>
<span class="sd"> Optional indices of the downsampling method, e.g. indices of the</span>
-<span class="sd"> maxima when using `torch.nn.functional.max_pool2d`. Useful for</span>
-<span class="sd"> upsampling later. If no indices are required to upsample, simply</span>
-<span class="sd"> return ``indices`` = `None`.</span>
+<span class="sd"> maxima when using :py:func:`torch.nn.functional.max_pool2d`. Useful</span>
+<span class="sd"> for upsampling later. If no indices are required to upsample,</span>
+<span class="sd"> simply return ``indices=None``.</span>
<span class="sd"> """</span>
<span class="c1"># the forward pass of the layers of the block</span>
@@ -389,57 +382,43 @@
<span class="k">def</span> <span class="nf">downsample</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">x</span><span class="p">):</span>
<span class="sd">"""Define the downsampling method.</span>
-<span class="sd"> The `~pysegcnn.core.layers.EncoderBlock.downsample` method should</span>
-<span class="sd"> implement the spatial pooling operation.</span>
+<span class="sd"> The :py:meth:`~pysegcnn.core.layers.EncoderBlock.downsample` method</span>
+<span class="sd"> should implement the spatial pooling operation.</span>
<span class="sd"> Use one of the following functions to downsample:</span>
-<span class="sd"> - `torch.nn.functional.max_pool2d`</span>
-<span class="sd"> - `torch.nn.functional.interpolate`</span>
+<span class="sd"> - :py:func:`torch.nn.functional.max_pool2d`</span>
+<span class="sd"> - :py:func:`torch.nn.functional.interpolate`</span>
-<span class="sd"> See `pysegcnn.core.layers.ConvBnReluMaxPool` for an example</span>
+<span class="sd"> See :py:class:`pysegcnn.core.layers.ConvBnReluMaxPool` for an example</span>
<span class="sd"> implementation.</span>
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
+<span class="sd"> x : :py:class:`torch.Tensor`, shape=(b, c, h, w)</span>
<span class="sd"> Input tensor, e.g. output of a convolutional block.</span>
<span class="sd"> Raises</span>
<span class="sd"> ------</span>
<span class="sd"> NotImplementedError</span>
-<span class="sd"> Raised if `pysegcnn.core.layers.EncoderBlock` is not inherited.</span>
+<span class="sd"> Raised if :py:class:`pysegcnn.core.layers.EncoderBlock` is not</span>
+<span class="sd"> inherited.</span>
<span class="sd"> Returns</span>
<span class="sd"> -------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
+<span class="sd"> x : :py:class:`torch.Tensor`, shape=(b, c, h, w)</span>
<span class="sd"> The spatially downsampled tensor.</span>
-<span class="sd"> indices : `torch.Tensor` or `None`</span>
+<span class="sd"> indices : :py:class:`torch.Tensor` or `None`</span>
<span class="sd"> Optional indices of the downsampling method, e.g. indices of the</span>
-<span class="sd"> maxima when using `torch.nn.functional.max_pool2d`. Useful for</span>
-<span class="sd"> upsampling later. If no indices are required to upsample, simply</span>
-<span class="sd"> return ``indices`` = `None`.</span>
+<span class="sd"> maxima when using :py:func:`torch.nn.functional.max_pool2d`. Useful</span>
+<span class="sd"> for upsampling later. If no indices are required to upsample,</span>
+<span class="sd"> simply return ``indices=None``.</span>
<span class="sd"> """</span>
<span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">(</span><span class="s1">'Implement the downsampling function.'</span><span class="p">)</span></div>
<div class="viewcode-block" id="DecoderBlock"><a class="viewcode-back" href="../../../source/generated/pysegcnn.core.layers.DecoderBlock.html#pysegcnn.core.layers.DecoderBlock">[docs]</a><span class="k">class</span> <span class="nc">DecoderBlock</span><span class="p">(</span><span class="n">Block</span><span class="p">):</span>
- <span class="sd">"""Block of a convolutional decoder.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> in_channels : `int`</span>
-<span class="sd"> Number of input channels.</span>
-<span class="sd"> out_channels : `int`</span>
-<span class="sd"> Number of output channels.</span>
-<span class="sd"> **kwargs: 'dict' [`str`]</span>
-<span class="sd"> Additional arguments passed to `pysegcnn.core.layers.Conv2dSame`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
+ <span class="sd">"""Block of a convolutional decoder."""</span>
<div class="viewcode-block" id="DecoderBlock.__init__"><a class="viewcode-back" href="../../../source/generated/pysegcnn.core.layers.DecoderBlock.html#pysegcnn.core.layers.DecoderBlock.__init__">[docs]</a> <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">in_channels</span><span class="p">,</span> <span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
<span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">in_channels</span><span class="p">,</span> <span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span></div>
@@ -449,21 +428,21 @@
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
+<span class="sd"> x : :py:class:`torch.Tensor`, shape=(b, c, h, w)</span>
<span class="sd"> Input tensor.</span>
-<span class="sd"> feature : `torch.Tensor`, shape=(batch, channel, height, width)</span>
-<span class="sd"> Intermediate output of a layer in the encoder.</span>
-<span class="sd"> If ``skip`` = True, ``feature`` is concatenated (along the channel</span>
-<span class="sd"> axis) to the output of the respective upsampling layer in the</span>
-<span class="sd"> decoder (skip connection).</span>
-<span class="sd"> indices : `torch.Tensor` or `None`</span>
+<span class="sd"> feature : :py:class:`torch.Tensor`, shape=(b, c, h, w)</span>
+<span class="sd"> Intermediate output of a layer in the encoder. If ``skip=True``,</span>
+<span class="sd"> ``feature`` is concatenated (along the channel axis) to the output</span>
+<span class="sd"> of the respective upsampling layer in the decoder (skip connection)</span>
+<span class="sd"> .</span>
+<span class="sd"> indices : :py:class:`torch.Tensor` or `None`</span>
<span class="sd"> Indices of the encoder downsampling method.</span>
<span class="sd"> skip : `bool`</span>
<span class="sd"> Whether to apply the skip connection.</span>
<span class="sd"> Returns</span>
<span class="sd"> -------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
+<span class="sd"> x : :py:class:`torch.Tensor`, shape=(b, c, h, w)</span>
<span class="sd"> Output of the decoder block.</span>
<span class="sd"> """</span>
@@ -484,35 +463,36 @@
<span class="k">def</span> <span class="nf">upsample</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">x</span><span class="p">,</span> <span class="n">feature</span><span class="p">,</span> <span class="n">indices</span><span class="p">):</span>
<span class="sd">"""Define the upsampling method.</span>
-<span class="sd"> The `~pysegcnn.core.layers.DecoderBlock.upsample` method should</span>
-<span class="sd"> implement the spatial upsampling operation.</span>
+<span class="sd"> The :py:meth:`~pysegcnn.core.layers.DecoderBlock.upsample` method</span>
+<span class="sd"> should implement the spatial upsampling operation.</span>
<span class="sd"> Use one of the following functions to upsample:</span>
-<span class="sd"> - `torch.nn.functional.max_unpool2d`</span>
-<span class="sd"> - `torch.nn.functional.interpolate`</span>
+<span class="sd"> - :py:func:`torch.nn.functional.max_unpool2d`</span>
+<span class="sd"> - :py:func:`torch.nn.functional.interpolate`</span>
-<span class="sd"> See `pysegcnn.core.layers.ConvBnReluMaxUnpool` or</span>
-<span class="sd"> `pysegcnn.core.layers.ConvBnReluUpsample` for an example</span>
+<span class="sd"> See :py:class:`pysegcnn.core.layers.ConvBnReluMaxUnpool` or</span>
+<span class="sd"> :py:class:`pysegcnn.core.layers.ConvBnReluUpsample` for an example</span>
<span class="sd"> implementation.</span>
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
+<span class="sd"> x : :py:class:`torch.Tensor`, shape=(b, c, h, w)</span>
<span class="sd"> Input tensor, e.g. output of a convolutional block.</span>
-<span class="sd"> feature : `torch.Tensor`, shape=(batch, channel, height, width)</span>
+<span class="sd"> feature : :py:class:`torch.Tensor`, shape=(b, c, h, w)</span>
<span class="sd"> Intermediate output of a layer in the encoder. Used to implement</span>
<span class="sd"> skip connections.</span>
-<span class="sd"> indices : `torch.Tensor` or `None`</span>
+<span class="sd"> indices : :py:class:`torch.Tensor` or `None`</span>
<span class="sd"> Indices of the encoder downsampling method.</span>
<span class="sd"> Raises</span>
<span class="sd"> ------</span>
<span class="sd"> NotImplementedError</span>
-<span class="sd"> Raised if `pysegcnn.core.layers.DecoderBlock` is not inherited.</span>
+<span class="sd"> Raised if :py:class:`pysegcnn.core.layers.DecoderBlock` is not</span>
+<span class="sd"> inherited.</span>
<span class="sd"> Returns</span>
<span class="sd"> -------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
+<span class="sd"> x : :py:class:`torch.Tensor`, shape=(b, c, h, w)</span>
<span class="sd"> The spatially upsampled tensor.</span>
<span class="sd"> """</span>
@@ -523,31 +503,50 @@
<span class="sd">"""Generic convolutional encoder.</span>
<span class="sd"> When instanciating an encoder-decoder architechure, ``filters`` should be</span>
-<span class="sd"> the same for `pysegcnn.core.layers.Encoder` and</span>
-<span class="sd"> `pysegcnn.core.layers.Decoder`.</span>
+<span class="sd"> the same for :py:class:`pysegcnn.core.layers.Encoder` and</span>
+<span class="sd"> :py:class:`pysegcnn.core.layers.Decoder`.</span>
-<span class="sd"> See `pysegcnn.core.models.UNet` for an example implementation.</span>
+<span class="sd"> See :py:class:`pysegcnn.core.models.UNet` for an example implementation.</span>
-<span class="sd"> Parameters</span>
+<span class="sd"> Attributes</span>
<span class="sd"> ----------</span>
-<span class="sd"> filters : `list` [`int`]</span>
-<span class="sd"> List of input channels to each convolutional block. The length of</span>
-<span class="sd"> ``filters`` determines the depth of the encoder. The first element of</span>
-<span class="sd"> ``filters`` has to be the number of channels of the input images.</span>
-<span class="sd"> block : `pysegcnn.core.layers.EncoderBlock`</span>
+<span class="sd"> features : :py:class:`numpy.ndarray`</span>
+<span class="sd"> Input channels to each convolutional block, i.e. ``filters``.</span>
+<span class="sd"> block : :py:class:`pysegcnn.core.layers.EncoderBlock`</span>
<span class="sd"> The convolutional block defining a layer in the encoder.</span>
-<span class="sd"> A subclass of `pysegcnn.core.layers.EncoderBlock`, e.g.</span>
-<span class="sd"> `pysegcnn.core.layers.ConvBnReluMaxPool`.</span>
-<span class="sd"> **kwargs: 'dict' [`str`]</span>
-<span class="sd"> Additional arguments passed to `pysegcnn.core.layers.Conv2dSame`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
+<span class="sd"> A subclass of :py:class:`pysegcnn.core.layers.EncoderBlock`, e.g.</span>
+<span class="sd"> :py:class:`pysegcnn.core.layers.ConvBnReluMaxPool`.</span>
+<span class="sd"> layers : :py:class:`torch.nn.ModuleList`</span>
+<span class="sd"> List of blocks in the encoder.</span>
+<span class="sd"> cache : `dict`</span>
+<span class="sd"> Intermediate encoder outputs. Dictionary with keys:</span>
+<span class="sd"> ``'feature'``</span>
+<span class="sd"> The intermediate encoder outputs (:py:class:`torch.Tensor`).</span>
+<span class="sd"> ``'indices'``</span>
+<span class="sd"> The indices of the max pooling layer, if required</span>
+<span class="sd"> (:py:class:`torch.Tensor`).</span>
<span class="sd"> """</span>
<div class="viewcode-block" id="Encoder.__init__"><a class="viewcode-back" href="../../../source/generated/pysegcnn.core.layers.Encoder.html#pysegcnn.core.layers.Encoder.__init__">[docs]</a> <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">filters</span><span class="p">,</span> <span class="n">block</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
+ <span class="sd">"""Initialize.</span>
+
+<span class="sd"> Parameters</span>
+<span class="sd"> ----------</span>
+<span class="sd"> filters : `list` [`int`]</span>
+<span class="sd"> List of input channels to each convolutional block. The length of</span>
+<span class="sd"> ``filters`` determines the depth of the encoder. The first element</span>
+<span class="sd"> of ``filters`` has to be the number of channels of the input</span>
+<span class="sd"> images.</span>
+<span class="sd"> block : :py:class:`pysegcnn.core.layers.EncoderBlock`</span>
+<span class="sd"> The convolutional block defining a layer in the encoder.</span>
+<span class="sd"> A subclass of :py:class:`pysegcnn.core.layers.EncoderBlock`, e.g.</span>
+<span class="sd"> :py:class:`pysegcnn.core.layers.ConvBnReluMaxPool`.</span>
+<span class="sd"> **kwargs: `dict` [`str`]</span>
+<span class="sd"> Additional arguments passed to</span>
+<span class="sd"> :py:class:`pysegcnn.core.layers.Conv2dSame`.</span>
+
+<span class="sd"> """</span>
<span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">()</span>
<span class="c1"># the number of filters for each block: the first element of filters</span>
@@ -573,22 +572,22 @@
<span class="sd">"""Forward pass of the encoder.</span>
<span class="sd"> Stores intermediate outputs in a dictionary. The keys of the dictionary</span>
-<span class="sd"> are the number of the network layers and the values are dictionaries</span>
+<span class="sd"> are the numbers of the network layers and the values are dictionaries</span>
<span class="sd"> with the following (key, value) pairs:</span>
-<span class="sd"> ``"feature"``</span>
-<span class="sd"> The intermediate encoder outputs (`torch.Tensor`).</span>
-<span class="sd"> ``"indices"``</span>
+<span class="sd"> ``'feature'``</span>
+<span class="sd"> The intermediate encoder outputs (:py:class:`torch.Tensor`).</span>
+<span class="sd"> ``'indices'``</span>
<span class="sd"> The indices of the max pooling layer, if required</span>
-<span class="sd"> (`torch.Tensor`).</span>
+<span class="sd"> (:py:class:`torch.Tensor`).</span>
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
+<span class="sd"> x : :py:class:`torch.Tensor`, shape=(b, c, h, w)</span>
<span class="sd"> Input image.</span>
<span class="sd"> Returns</span>
<span class="sd"> -------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
+<span class="sd"> x : :py:class:`torch.Tensor`, shape=(b, c, h, w)</span>
<span class="sd"> Output of the encoder.</span>
<span class="sd"> """</span>
@@ -609,33 +608,47 @@
<span class="sd">"""Generic convolutional decoder.</span>
<span class="sd"> When instanciating an encoder-decoder architechure, ``filters`` should be</span>
-<span class="sd"> the same for `pysegcnn.core.layers.Encoder` and</span>
-<span class="sd"> `pysegcnn.core.layers.Decoder`.</span>
+<span class="sd"> the same for :py:class:`pysegcnn.core.layers.Encoder` and</span>
+<span class="sd"> :py:class:`pysegcnn.core.layers.Decoder`.</span>
-<span class="sd"> See `pysegcnn.core.models.UNet` for an example implementation.</span>
+<span class="sd"> See :py:class:`pysegcnn.core.models.UNet` for an example implementation.</span>
-<span class="sd"> Parameters</span>
+<span class="sd"> Attributes</span>
<span class="sd"> ----------</span>
-<span class="sd"> filters : `list` [`int`]</span>
-<span class="sd"> List of input channels to each convolutional block. The length of</span>
-<span class="sd"> ``filters`` determines the depth of the decoder. The first element of</span>
-<span class="sd"> ``filters`` has to be the number of channels of the input images.</span>
-<span class="sd"> block : `pysegcnn.core.layers.DecoderBlock`</span>
+<span class="sd"> features : :py:class:`numpy.ndarray`</span>
+<span class="sd"> Input channels to each convolutional block, i.e. ``filters``.</span>
+<span class="sd"> block : :py:class:`pysegcnn.core.layers.EncoderBlock`</span>
<span class="sd"> The convolutional block defining a layer in the decoder.</span>
-<span class="sd"> A subclass of `pysegcnn.core.layers.DecoderBlock`, e.g.</span>
-<span class="sd"> `pysegcnn.core.layers.ConvBnReluMaxUnpool`.</span>
+<span class="sd"> A subclass of :py:class:`pysegcnn.core.layers.EncoderBlock`, e.g.</span>
+<span class="sd"> :py:class:`pysegcnn.core.layers.ConvBnReluMaxPool`.</span>
<span class="sd"> skip : `bool`</span>
<span class="sd"> Whether to apply skip connections from the encoder to the decoder.</span>
-<span class="sd"> **kwargs: 'dict' [`str`]</span>
-<span class="sd"> Additional arguments passed to `pysegcnn.core.layers.Conv2dSame`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
+<span class="sd"> layers : :py:class:`torch.nn.ModuleList`</span>
+<span class="sd"> List of blocks in the decoder.</span>
<span class="sd"> """</span>
<div class="viewcode-block" id="Decoder.__init__"><a class="viewcode-back" href="../../../source/generated/pysegcnn.core.layers.Decoder.html#pysegcnn.core.layers.Decoder.__init__">[docs]</a> <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">filters</span><span class="p">,</span> <span class="n">block</span><span class="p">,</span> <span class="n">skip</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
+ <span class="sd">"""Initialize.</span>
+
+<span class="sd"> Parameters</span>
+<span class="sd"> ----------</span>
+<span class="sd"> filters : `list` [`int`]</span>
+<span class="sd"> List of input channels to each convolutional block. The length of</span>
+<span class="sd"> ``filters`` determines the depth of the decoder. The first element</span>
+<span class="sd"> of ``filters`` has to be the number of channels of the input</span>
+<span class="sd"> images.</span>
+<span class="sd"> block : :py:class:`pysegcnn.core.layers.DecoderBlock`</span>
+<span class="sd"> The convolutional block defining a layer in the decoder.</span>
+<span class="sd"> A subclass of :py:class:`pysegcnn.core.layers.DecoderBlock`, e.g.</span>
+<span class="sd"> :py:class:`pysegcnn.core.layers.ConvBnReluMaxUnpool`.</span>
+<span class="sd"> skip : `bool`, optional</span>
+<span class="sd"> Whether to apply skip connections from the encoder to the decoder.</span>
+<span class="sd"> **kwargs: `dict` [`str`]</span>
+<span class="sd"> Additional arguments passed to</span>
+<span class="sd"> :py:class:`pysegcnn.core.layers.Conv2dSame`.</span>
+
+<span class="sd"> """</span>
<span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">()</span>
<span class="c1"># the block of operations defining a layer in the decoder</span>
@@ -670,20 +683,22 @@
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
+<span class="sd"> x : :py:class:`torch.Tensor`, shape=(b, c, h, w)</span>
<span class="sd"> Output of the encoder.</span>
<span class="sd"> enc_cache : `dict` [`dict`]</span>
<span class="sd"> Cache dictionary. The keys of the dictionary are the number of the</span>
<span class="sd"> network layers and the values are dictionaries with the following</span>
<span class="sd"> (key, value) pairs:</span>
-<span class="sd"> ``"feature"``</span>
-<span class="sd"> The intermediate encoder outputs (`torch.Tensor`).</span>
-<span class="sd"> ``"indices"``</span>
-<span class="sd"> The indices of the max pooling layer (`torch.Tensor`).</span>
+<span class="sd"> ``'feature'``</span>
+<span class="sd"> The intermediate encoder outputs</span>
+<span class="sd"> (:py:class:`torch.Tensor`).</span>
+<span class="sd"> ``'indices'``</span>
+<span class="sd"> The indices of the max pooling layer</span>
+<span class="sd"> (:py:class:`torch.Tensor`).</span>
<span class="sd"> Returns</span>
<span class="sd"> -------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
+<span class="sd"> x : :py:class:`torch.Tensor`, shape=(b, c, h, w)</span>
<span class="sd"> Output of the decoder.</span>
<span class="sd"> """</span>
@@ -702,23 +717,7 @@
<div class="viewcode-block" id="ConvBnReluMaxPool"><a class="viewcode-back" href="../../../source/generated/pysegcnn.core.layers.ConvBnReluMaxPool.html#pysegcnn.core.layers.ConvBnReluMaxPool">[docs]</a><span class="k">class</span> <span class="nc">ConvBnReluMaxPool</span><span class="p">(</span><span class="n">EncoderBlock</span><span class="p">):</span>
- <span class="sd">"""Block of convolution, batchnorm, relu and 2x2 max pool.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> in_channels : `int`</span>
-<span class="sd"> Number of input channels.</span>
-<span class="sd"> out_channels : `int`</span>
-<span class="sd"> Number of output channels.</span>
-<span class="sd"> **kwargs: 'dict' [`str`]</span>
-<span class="sd"> Additional keyword arguments passed to</span>
-<span class="sd"> `pysegcnn.core.layers.Conv2dSame`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
+ <span class="sd">"""Block of convolution, batchnorm, relu and 2x2 max pool."""</span>
<div class="viewcode-block" id="ConvBnReluMaxPool.__init__"><a class="viewcode-back" href="../../../source/generated/pysegcnn.core.layers.ConvBnReluMaxPool.html#pysegcnn.core.layers.ConvBnReluMaxPool.__init__">[docs]</a> <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">in_channels</span><span class="p">,</span> <span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
<span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">in_channels</span><span class="p">,</span> <span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span></div>
@@ -728,29 +727,30 @@
<span class="sd"> Returns</span>
<span class="sd"> -------</span>
-<span class="sd"> layers : `torch.nn.Sequential` [`torch.nn.Module`]</span>
-<span class="sd"> An instance of `torch.nn.Sequential` containing the sequence</span>
-<span class="sd"> of convolution, batchnorm and relu layer (`torch.nn.Module`)</span>
-<span class="sd"> instances.</span>
+<span class="sd"> layers : :py:class:`torch.nn.Sequential` [:py:class:`torch.nn.Module`]</span>
+<span class="sd"> An instance of :py:class:`torch.nn.Sequential` containing the</span>
+<span class="sd"> sequence of convolution, batchnorm and relu layer</span>
+<span class="sd"> (:py:class:`torch.nn.Module`) instances.</span>
<span class="sd"> """</span>
<span class="k">return</span> <span class="n">conv_bn_relu</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">in_channels</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="bp">self</span><span class="o">.</span><span class="n">kwargs</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">downsample</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">x</span><span class="p">):</span>
- <span class="sd">"""2x2 max pooling layer, `torch.nn.functional.max_pool2d`.</span>
+ <span class="sd">"""2x2 max pooling layer.</span>
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
+<span class="sd"> x : :py:class:`torch.Tensor`, shape=(b, c, h, w)</span>
<span class="sd"> Input tensor.</span>
<span class="sd"> Returns</span>
<span class="sd"> -------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height // 2, width // 2)</span>
+<span class="sd"> x : :py:class:`torch.Tensor`, shape=(b, c, h // 2, w // 2)</span>
<span class="sd"> The 2x2 max pooled tensor.</span>
-<span class="sd"> indices : `torch.Tensor` or `None`</span>
+<span class="sd"> indices : :py:class:`torch.Tensor` or `None`</span>
<span class="sd"> The indices of the maxima. Useful for upsampling with</span>
-<span class="sd"> `torch.nn.functional.max_unpool2d`.</span>
+<span class="sd"> :py:func:`torch.nn.functional.max_unpool2d`.</span>
+
<span class="sd"> """</span>
<span class="n">x</span><span class="p">,</span> <span class="n">indices</span> <span class="o">=</span> <span class="n">F</span><span class="o">.</span><span class="n">max_pool2d</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">kernel_size</span><span class="o">=</span><span class="mi">2</span><span class="p">,</span> <span class="n">return_indices</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
<span class="k">return</span> <span class="n">x</span><span class="p">,</span> <span class="n">indices</span>
@@ -768,23 +768,7 @@
<div class="viewcode-block" id="ConvBnReluMaxUnpool"><a class="viewcode-back" href="../../../source/generated/pysegcnn.core.layers.ConvBnReluMaxUnpool.html#pysegcnn.core.layers.ConvBnReluMaxUnpool">[docs]</a><span class="k">class</span> <span class="nc">ConvBnReluMaxUnpool</span><span class="p">(</span><span class="n">DecoderBlock</span><span class="p">):</span>
- <span class="sd">"""Block of convolution, batchnorm, relu and 2x2 max unpool.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> in_channels : `int`</span>
-<span class="sd"> Number of input channels.</span>
-<span class="sd"> out_channels : `int`</span>
-<span class="sd"> Number of output channels</span>
-<span class="sd"> **kwargs: 'dict' [`str`]</span>
-<span class="sd"> Additional keyword arguments passed to</span>
-<span class="sd"> `pysegcnn.core.layers.Conv2dSame`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
+ <span class="sd">"""Block of convolution, batchnorm, relu and 2x2 max unpool."""</span>
<div class="viewcode-block" id="ConvBnReluMaxUnpool.__init__"><a class="viewcode-back" href="../../../source/generated/pysegcnn.core.layers.ConvBnReluMaxUnpool.html#pysegcnn.core.layers.ConvBnReluMaxUnpool.__init__">[docs]</a> <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">in_channels</span><span class="p">,</span> <span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
<span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">in_channels</span><span class="p">,</span> <span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span></div>
@@ -794,10 +778,10 @@
<span class="sd"> Returns</span>
<span class="sd"> -------</span>
-<span class="sd"> layers : `torch.nn.Sequential` [`torch.nn.Module`]</span>
-<span class="sd"> An instance of `torch.nn.Sequential` containing the sequence</span>
-<span class="sd"> of convolution, batchnorm and relu layer (`torch.nn.Module`)</span>
-<span class="sd"> instances.</span>
+<span class="sd"> layers : :py:class:`torch.nn.Sequential` [:py:class:`torch.nn.Module`]</span>
+<span class="sd"> An instance of :py:class:`torch.nn.Sequential` containing the</span>
+<span class="sd"> sequence of convolution, batchnorm and relu layer</span>
+<span class="sd"> (:py:class:`torch.nn.Module`) instances.</span>
<span class="sd"> """</span>
<span class="k">return</span> <span class="n">conv_bn_relu</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">in_channels</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="bp">self</span><span class="o">.</span><span class="n">kwargs</span><span class="p">)</span>
@@ -807,18 +791,18 @@
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
+<span class="sd"> x : :py:class:`torch.Tensor`, shape=(b, c, h, w)</span>
<span class="sd"> Input tensor.</span>
-<span class="sd"> feature : `torch.Tensor`, shape=(batch, channel, height, width)</span>
+<span class="sd"> feature : :py:class:`torch.Tensor`, shape=(b, c, h, w)</span>
<span class="sd"> Intermediate output of a layer in the encoder. Used to determine</span>
<span class="sd"> the output shape of the upsampling operation.</span>
-<span class="sd"> indices : `torch.Tensor`</span>
+<span class="sd"> indices : :py:class:`torch.Tensor`</span>
<span class="sd"> The indices of the maxima of the max pooling operation</span>
-<span class="sd"> (as returned by `torch.nn.functional.max_pool2d`).</span>
+<span class="sd"> (as returned by :py:func:`torch.nn.functional.max_pool2d`).</span>
<span class="sd"> Returns</span>
<span class="sd"> -------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height * 2, width * 2)</span>
+<span class="sd"> x : :py:class:`torch.Tensor`, shape=(b, c, h * 2, w * 2)</span>
<span class="sd"> The 2x2 max unpooled tensor.</span>
<span class="sd"> """</span>
@@ -838,22 +822,7 @@
<div class="viewcode-block" id="ConvBnReluUpsample"><a class="viewcode-back" href="../../../source/generated/pysegcnn.core.layers.ConvBnReluUpsample.html#pysegcnn.core.layers.ConvBnReluUpsample">[docs]</a><span class="k">class</span> <span class="nc">ConvBnReluUpsample</span><span class="p">(</span><span class="n">DecoderBlock</span><span class="p">):</span>
- <span class="sd">"""Block of convolution, batchnorm, relu and nearest neighbor upsampling.</span>
-
-<span class="sd"> Parameters</span>
-<span class="sd"> ----------</span>
-<span class="sd"> in_channels : `int`</span>
-<span class="sd"> Number of input channels.</span>
-<span class="sd"> out_channels : `int`</span>
-<span class="sd"> Number of output channels</span>
-<span class="sd"> **kwargs: 'dict' [`str`]</span>
-<span class="sd"> Additional arguments passed to `pysegcnn.core.layers.Conv2dSame`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
-
-<span class="sd"> """</span>
+ <span class="sd">"""Block of convolution, batchnorm, relu and nearest neighbor upsample."""</span>
<div class="viewcode-block" id="ConvBnReluUpsample.__init__"><a class="viewcode-back" href="../../../source/generated/pysegcnn.core.layers.ConvBnReluUpsample.html#pysegcnn.core.layers.ConvBnReluUpsample.__init__">[docs]</a> <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">in_channels</span><span class="p">,</span> <span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
<span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">in_channels</span><span class="p">,</span> <span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span></div>
@@ -863,10 +832,10 @@
<span class="sd"> Returns</span>
<span class="sd"> -------</span>
-<span class="sd"> layers : `torch.nn.Sequential` [`torch.nn.Module`]</span>
-<span class="sd"> An instance of `torch.nn.Sequential` containing the sequence</span>
-<span class="sd"> of convolution, batchnorm and relu layer (`torch.nn.Module`)</span>
-<span class="sd"> instances.</span>
+<span class="sd"> layers : :py:class:`torch.nn.Sequential` [:py:class:`torch.nn.Module`]</span>
+<span class="sd"> An instance of :py:class:`torch.nn.Sequential` containing the</span>
+<span class="sd"> sequence of convolution, batchnorm and relu layer</span>
+<span class="sd"> (:py:class:`torch.nn.Module`) instances.</span>
<span class="sd"> """</span>
<span class="k">return</span> <span class="n">conv_bn_relu</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">in_channels</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">out_channels</span><span class="p">,</span> <span class="o">**</span><span class="bp">self</span><span class="o">.</span><span class="n">kwargs</span><span class="p">)</span>
@@ -876,19 +845,19 @@
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
+<span class="sd"> x : :py:class:`torch.Tensor`, shape=(b, c, h, w)</span>
<span class="sd"> Input tensor.</span>
-<span class="sd"> feature : `torch.Tensor`, shape=(batch, channel, height, width)</span>
+<span class="sd"> feature : :py:class:`torch.Tensor`, shape=(b, c, h, w)</span>
<span class="sd"> Intermediate output of a layer in the encoder. Used to determine</span>
<span class="sd"> the output shape of the upsampling operation.</span>
<span class="sd"> indices : `None`, optional</span>
<span class="sd"> The indices of the maxima of the max pooling operation</span>
-<span class="sd"> (as returned by `torch.nn.functional.max_pool2d`). Not required by</span>
-<span class="sd"> this upsampling method.</span>
+<span class="sd"> (as returned by :py:func:`torch.nn.functional.max_pool2d`).</span>
+<span class="sd"> Not required by this upsampling method.</span>
<span class="sd"> Returns</span>
<span class="sd"> -------</span>
-<span class="sd"> x : `torch.Tensor`, shape=(batch, channel, height, width)</span>
+<span class="sd"> x : :py:class:`torch.Tensor`, shape=(b, c, h, w)</span>
<span class="sd"> The 2x2 upsampled tensor.</span>
<span class="sd"> """</span>
diff --git a/docs/_build/html/_modules/pysegcnn/core/models.html b/docs/_build/html/_modules/pysegcnn/core/models.html
index ed55a3991df1d9525761ae41aabefb7b395024a7..4e951b1583dc86a5a4d0488d06509f8f46e1c8f6 100644
--- a/docs/_build/html/_modules/pysegcnn/core/models.html
+++ b/docs/_build/html/_modules/pysegcnn/core/models.html
@@ -187,15 +187,17 @@
<span class="sd">"""Generic Network class.</span>
<span class="sd"> The base class for each model. If you want to implement a new model,</span>
-<span class="sd"> inherit the ``~pysegcnn.core.models.Network`` class.</span>
+<span class="sd"> inherit the :py:class:`pysegcnn.core.models.Network` class.</span>
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
+<span class="sd"> Attributes</span>
+<span class="sd"> ----------</span>
+<span class="sd"> state_file : `str` or `None` or :py:class:`pathlib.Path`</span>
+<span class="sd"> The model state file, where the model parameters are saved.</span>
<span class="sd"> """</span>
<div class="viewcode-block" id="Network.__init__"><a class="viewcode-back" href="../../../source/generated/pysegcnn.core.models.Network.html#pysegcnn.core.models.Network.__init__">[docs]</a> <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
+ <span class="sd">"""Initialize."""</span>
<span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">()</span>
<span class="c1"># initialize state file</span>
@@ -239,9 +241,9 @@
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
-<span class="sd"> state_file : `str` or `pathlib.Path`</span>
+<span class="sd"> state_file : `str` or :py:class:`pathlib.Path`</span>
<span class="sd"> Path to save the model state.</span>
-<span class="sd"> optimizer : `torch.optim.Optimizer`</span>
+<span class="sd"> optimizer : :py:class:`torch.optim.Optimizer`</span>
<span class="sd"> The optimizer used to train the model.</span>
<span class="sd"> bands : `list` [`str`] or `None`, optional</span>
<span class="sd"> List of bands the model is trained with. The default is None.</span>
@@ -252,7 +254,7 @@
<span class="sd"> Returns</span>
<span class="sd"> -------</span>
<span class="sd"> model_state : `dict`</span>
-<span class="sd"> A dictionary containing the model and optimizer state</span>
+<span class="sd"> A dictionary containing the model and optimizer state.</span>
<span class="sd"> """</span>
<span class="c1"># check if the output path exists and if not, create it</span>
@@ -306,7 +308,7 @@
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
-<span class="sd"> state_file : `str` or `pathlib.Path`</span>
+<span class="sd"> state_file : `str` or :py:class:`pathlib.Path`</span>
<span class="sd"> The model state file. Model state files are stored in</span>
<span class="sd"> pysegcnn/main/_models.</span>
@@ -317,13 +319,13 @@
<span class="sd"> Returns</span>
<span class="sd"> -------</span>
-<span class="sd"> model : `pysegcnn.core.models.Network`</span>
+<span class="sd"> model : :py:class:`pysegcnn.core.models.Network`</span>
<span class="sd"> The pretrained model.</span>
-<span class="sd"> optimizer : `torch.optim.Optimizer`</span>
+<span class="sd"> optimizer : :py:class:`torch.optim.Optimizer`</span>
<span class="sd"> The optimizer used to train the model.</span>
-<span class="sd"> model_state : '`dict`</span>
+<span class="sd"> model_state : `dict`</span>
<span class="sd"> A dictionary containing the model and optimizer state, as</span>
-<span class="sd"> constructed by `~pysegcnn.core.Network.save`.</span>
+<span class="sd"> constructed by :py:meth:`~pysegcnn.core.Network.save`.</span>
<span class="sd"> """</span>
<span class="c1"># load the pretrained model</span>
@@ -366,7 +368,7 @@
<span class="sd"> Returns</span>
<span class="sd"> -------</span>
-<span class="sd"> state_file : `pathlib.Path` or `None`</span>
+<span class="sd"> state_file : :py:class:`pathlib.Path` or `None`</span>
<span class="sd"> The model state file.</span>
<span class="sd"> """</span>
@@ -374,36 +376,58 @@
<div class="viewcode-block" id="UNet"><a class="viewcode-back" href="../../../source/generated/pysegcnn.core.models.UNet.html#pysegcnn.core.models.UNet">[docs]</a><span class="k">class</span> <span class="nc">UNet</span><span class="p">(</span><span class="n">Network</span><span class="p">):</span>
- <span class="sd">"""A PyTorch implementation of `U-Net`_.</span>
+ <span class="sd">"""A slightly modified implementation of `U-Net`_ in PyTorch.</span>
+
+<span class="sd"> .. important::</span>
-<span class="sd"> Slightly modified version of U-Net:</span>
-<span class="sd"> - each convolution is followed by a batch normalization layer</span>
-<span class="sd"> - the upsampling is implemented by a 2x2 max unpooling operation</span>
+<span class="sd"> - Each convolution is followed by a batch normalization layer</span>
+<span class="sd"> - Upsampling is implemented by a 2x2 max unpooling operation</span>
<span class="sd"> .. _U-Net:</span>
<span class="sd"> https://arxiv.org/abs/1505.04597</span>
-<span class="sd"> Parameters</span>
+<span class="sd"> Attributes</span>
<span class="sd"> ----------</span>
<span class="sd"> in_channels : `int`</span>
<span class="sd"> Number of channels of the input images.</span>
<span class="sd"> nclasses : `int`</span>
<span class="sd"> Number of classes.</span>
-<span class="sd"> filters : `list` [`int`]</span>
+<span class="sd"> kwargs : `dict` [`str`]</span>
+<span class="sd"> Additional keyword arguments passed to</span>
+<span class="sd"> :py:class:`pysegcnn.core.layers.Conv2dSame`.</span>
+<span class="sd"> nfilters : `list` [`int`]</span>
<span class="sd"> List of input channels to each convolutional block.</span>
<span class="sd"> skip : `bool`</span>
<span class="sd"> Whether to apply skip connections from the encoder to the decoder.</span>
-<span class="sd"> **kwargs: 'dict' [`str`]</span>
-<span class="sd"> Additional keyword arguments passed to</span>
-<span class="sd"> `pysegcnn.core.layers.Conv2dSame`.</span>
-
-<span class="sd"> Returns</span>
-<span class="sd"> -------</span>
-<span class="sd"> None.</span>
+<span class="sd"> epoch : `int`</span>
+<span class="sd"> Number of epochs the model was trained.</span>
+<span class="sd"> encoder : :py:class:`pysegcnn.core.layers.Encoder`</span>
+<span class="sd"> The convolutional encoder.</span>
+<span class="sd"> decoder : :py:class:`pysegcnn.core.layers.Decoder`</span>
+<span class="sd"> The convolutional decoder.</span>
+<span class="sd"> classifier : :py:class:`pysegcnn.core.layers.Conv2dSame`</span>
+<span class="sd"> The classification layer, a 1x1 convolution.</span>
<span class="sd"> """</span>
<div class="viewcode-block" id="UNet.__init__"><a class="viewcode-back" href="../../../source/generated/pysegcnn.core.models.UNet.html#pysegcnn.core.models.UNet.__init__">[docs]</a> <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">in_channels</span><span class="p">,</span> <span class="n">nclasses</span><span class="p">,</span> <span class="n">filters</span><span class="p">,</span> <span class="n">skip</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
+ <span class="sd">"""Initialize.</span>
+
+<span class="sd"> Parameters</span>
+<span class="sd"> ----------</span>
+<span class="sd"> in_channels : `int`</span>
+<span class="sd"> Number of channels of the input images.</span>
+<span class="sd"> nclasses : `int`</span>
+<span class="sd"> Number of classes.</span>
+<span class="sd"> filters : `list` [`int`]</span>
+<span class="sd"> List of input channels to each convolutional block.</span>
+<span class="sd"> skip : `bool`</span>
+<span class="sd"> Whether to apply skip connections from the encoder to the decoder.</span>
+<span class="sd"> **kwargs: `dict` [`str`]</span>
+<span class="sd"> Additional keyword arguments passed to</span>
+<span class="sd"> :py:class:`pysegcnn.core.layers.Conv2dSame`.</span>
+
+<span class="sd"> """</span>
<span class="nb">super</span><span class="p">()</span><span class="o">.</span><span class="fm">__init__</span><span class="p">()</span>
<span class="c1"># number of input channels</span>
diff --git a/docs/_build/html/_sources/modules.rst.txt b/docs/_build/html/_sources/modules.rst.txt
deleted file mode 100644
index 383d07e61cdeb0228f6aa0a14dc4a376932a8a66..0000000000000000000000000000000000000000
--- a/docs/_build/html/_sources/modules.rst.txt
+++ /dev/null
@@ -1,10 +0,0 @@
-Modules
-=======
-
-.. autosummary::
- :toctree: modules
-
- pysegcnn.core.cli
- pysegcnn.core.constants
- pysegcnn.core.dataset
-
diff --git a/docs/_build/html/_sources/modules/pysegcnn.core.cli.rst.txt b/docs/_build/html/_sources/modules/pysegcnn.core.cli.rst.txt
deleted file mode 100644
index 3a5c0e3bb6b6d5c842eb2ea90dadd368efdb5f27..0000000000000000000000000000000000000000
--- a/docs/_build/html/_sources/modules/pysegcnn.core.cli.rst.txt
+++ /dev/null
@@ -1,29 +0,0 @@
-pysegcnn.core.cli
-=================
-
-.. automodule:: pysegcnn.core.cli
-
-
-
-
-
-
-
- .. rubric:: Functions
-
- .. autosummary::
-
- structure_parser
-
-
-
-
-
-
-
-
-
-
-
-
-
diff --git a/docs/_build/html/_sources/modules/pysegcnn.core.constants.rst.txt b/docs/_build/html/_sources/modules/pysegcnn.core.constants.rst.txt
deleted file mode 100644
index abb68f18fa7b316f5dad1bf7a8cfb48937cb9cf3..0000000000000000000000000000000000000000
--- a/docs/_build/html/_sources/modules/pysegcnn.core.constants.rst.txt
+++ /dev/null
@@ -1,34 +0,0 @@
-pysegcnn.core.constants
-=======================
-
-.. automodule:: pysegcnn.core.constants
-
-
-
-
-
-
-
-
-
-
-
- .. rubric:: Classes
-
- .. autosummary::
-
- Cloud95Labels
- Label
- Landsat8
- ProSnowLabels
- Sentinel2
- SparcsLabels
-
-
-
-
-
-
-
-
-
diff --git a/docs/_build/html/_sources/modules/pysegcnn.core.dataset.rst.txt b/docs/_build/html/_sources/modules/pysegcnn.core.dataset.rst.txt
deleted file mode 100644
index 2586d7412cd824ebed8c5cb1137f4b83a3147ef7..0000000000000000000000000000000000000000
--- a/docs/_build/html/_sources/modules/pysegcnn.core.dataset.rst.txt
+++ /dev/null
@@ -1,36 +0,0 @@
-pysegcnn.core.dataset
-=====================
-
-.. automodule:: pysegcnn.core.dataset
-
-
-
-
-
-
-
-
-
-
-
- .. rubric:: Classes
-
- .. autosummary::
-
- Cloud95Dataset
- ImageDataset
- ProSnowDataset
- ProSnowGarmisch
- ProSnowObergurgl
- SparcsDataset
- StandardEoDataset
- SupportedDatasets
-
-
-
-
-
-
-
-
-
diff --git a/docs/_build/html/_sources/readme.rst.txt b/docs/_build/html/_sources/readme.rst.txt
deleted file mode 100644
index 35b29b40bef7bea95bc65074292883308f4e6271..0000000000000000000000000000000000000000
--- a/docs/_build/html/_sources/readme.rst.txt
+++ /dev/null
@@ -1,4 +0,0 @@
-Test
-====
-
-.. mdinclude:: ../README.md
\ No newline at end of file
diff --git a/docs/_build/html/_sources/source/api.rst.txt b/docs/_build/html/_sources/source/api.rst.txt
index 4de52de54bd3cdf346bfdf3d35e6bef1ce7643ab..6d2aabc35a3d4889c14b817a6580618244188067 100644
--- a/docs/_build/html/_sources/source/api.rst.txt
+++ b/docs/_build/html/_sources/source/api.rst.txt
@@ -13,9 +13,9 @@ Dataset
Custom dataset classes compliant to the PyTorch `standard <https://pytorch.org/docs/stable/data.html>`_.
-Image Dataset
-^^^^^^^^^^^^^
-Generic class to implement `custom datasets <https://pytorch.org/tutorials/beginner/data_loading_tutorial.html>`_.
+Generic classes
+^^^^^^^^^^^^^^^
+Generic class to implement a `custom dataset <https://pytorch.org/tutorials/beginner/data_loading_tutorial.html>`_.
.. autosummary::
:toctree: generated/
@@ -23,11 +23,21 @@ Generic class to implement `custom datasets <https://pytorch.org/tutorials/begin
core.dataset.ImageDataset
-Supported datasets
-^^^^^^^^^^^^^^^^^^
+Generic class to implement a custom dataset following a standard directory
+structure.
+
+.. autosummary::
+ :toctree: generated/
+ :nosignatures:
+
+ core.dataset.StandardEoDataset
+
-The following open-source spaceborne multispectral image datasets are supported
-out-of-the-box:
+Specific classes
+^^^^^^^^^^^^^^^^
+
+Specific classes for some open-source image datasets. Currently, the following
+spaceborne multispectral image datasets are supported out-of-the-box:
.. autosummary::
:toctree: generated/
@@ -36,7 +46,6 @@ out-of-the-box:
core.dataset.SparcsDataset
core.dataset.Cloud95Dataset
-
Models
------
@@ -81,6 +90,9 @@ Neural Networks
core.models.UNet
+..
+ Links:
+
.. _Sparcs:
https://www.usgs.gov/land-resources/nli/landsat/spatial-procedures-automated-removal-cloud-and-shadow-sparcs-validation
@@ -92,3 +104,12 @@ Neural Networks
.. _U-Net:
https://arxiv.org/abs/1505.04597
+
+.. _Landsat-8:
+ https://www.usgs.gov/land-resources/nli/landsat/landsat-8?qt-science_support_page_related_con=0#qt-science_support_page_related_con
+
+.. _Sentinel-2:
+ https://sentinel.esa.int/web/sentinel/missions/sentinel-2
+
+.. _Hughes & Hayes (2014):
+ https://www.mdpi.com/2072-4292/6/6/4907
diff --git a/docs/_build/html/_sources/source/core.rst.txt b/docs/_build/html/_sources/source/core.rst.txt
deleted file mode 100644
index e57f944a0b43cc53179c8ac6638aebb1a32920a2..0000000000000000000000000000000000000000
--- a/docs/_build/html/_sources/source/core.rst.txt
+++ /dev/null
@@ -1,109 +0,0 @@
-core package
-============
-
-Submodules
-----------
-
-core.cli module
----------------
-
-.. automodule:: core.cli
- :members:
- :undoc-members:
- :show-inheritance:
-
-core.constants module
----------------------
-
-.. automodule:: core.constants
- :members:
- :undoc-members:
- :show-inheritance:
-
-core.dataset module
--------------------
-
-.. automodule:: core.dataset
- :members:
- :undoc-members:
- :show-inheritance:
-
-core.graphics module
---------------------
-
-.. automodule:: core.graphics
- :members:
- :undoc-members:
- :show-inheritance:
-
-core.layers module
-------------------
-
-.. automodule:: core.layers
- :members:
- :undoc-members:
- :show-inheritance:
-
-core.logging module
--------------------
-
-.. automodule:: core.logging
- :members:
- :undoc-members:
- :show-inheritance:
-
-core.models module
-------------------
-
-.. automodule:: core.models
- :members:
- :undoc-members:
- :show-inheritance:
-
-core.predict module
--------------------
-
-.. automodule:: core.predict
- :members:
- :undoc-members:
- :show-inheritance:
-
-core.split module
------------------
-
-.. automodule:: core.split
- :members:
- :undoc-members:
- :show-inheritance:
-
-core.trainer module
--------------------
-
-.. automodule:: core.trainer
- :members:
- :undoc-members:
- :show-inheritance:
-
-core.transforms module
-----------------------
-
-.. automodule:: core.transforms
- :members:
- :undoc-members:
- :show-inheritance:
-
-core.utils module
------------------
-
-.. automodule:: core.utils
- :members:
- :undoc-members:
- :show-inheritance:
-
-Module contents
----------------
-
-.. automodule:: core
- :members:
- :undoc-members:
- :show-inheritance:
diff --git a/docs/_build/html/_sources/source/generated/pysegcnn.core.dataset.Cloud95Dataset.rst.txt b/docs/_build/html/_sources/source/generated/pysegcnn.core.dataset.Cloud95Dataset.rst.txt
index 0b8650621242af8de39fc4bf54f43f371b295a94..a1810656ee6dab330ce851fe7a3320f4396a65d0 100644
--- a/docs/_build/html/_sources/source/generated/pysegcnn.core.dataset.Cloud95Dataset.rst.txt
+++ b/docs/_build/html/_sources/source/generated/pysegcnn.core.dataset.Cloud95Dataset.rst.txt
@@ -1,4 +1,4 @@
-pysegcnn.core.dataset.Cloud95Dataset
+pysegcnn.core.dataset.Cloud95Dataset
====================================
.. currentmodule:: pysegcnn.core.dataset
diff --git a/docs/_build/html/_sources/source/generated/pysegcnn.core.dataset.ImageDataset.rst.txt b/docs/_build/html/_sources/source/generated/pysegcnn.core.dataset.ImageDataset.rst.txt
index c04ed1f96ea2d230023e87a20bebeb31c0fe1050..31c36ff73f998226baf6cf9c48daeaea74404f94 100644
--- a/docs/_build/html/_sources/source/generated/pysegcnn.core.dataset.ImageDataset.rst.txt
+++ b/docs/_build/html/_sources/source/generated/pysegcnn.core.dataset.ImageDataset.rst.txt
@@ -1,4 +1,4 @@
-pysegcnn.core.dataset.ImageDataset
+pysegcnn.core.dataset.ImageDataset
==================================
.. currentmodule:: pysegcnn.core.dataset
diff --git a/docs/_build/html/_sources/source/generated/pysegcnn.core.dataset.SparcsDataset.rst.txt b/docs/_build/html/_sources/source/generated/pysegcnn.core.dataset.SparcsDataset.rst.txt
index d5c9fe46ca1985f87d71a260f0a9168ce193c5c4..cb357b5f003e77960920e99d3fe0ec24ec014b46 100644
--- a/docs/_build/html/_sources/source/generated/pysegcnn.core.dataset.SparcsDataset.rst.txt
+++ b/docs/_build/html/_sources/source/generated/pysegcnn.core.dataset.SparcsDataset.rst.txt
@@ -1,4 +1,4 @@
-pysegcnn.core.dataset.SparcsDataset
+pysegcnn.core.dataset.SparcsDataset
===================================
.. currentmodule:: pysegcnn.core.dataset
diff --git a/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.Block.rst.txt b/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.Block.rst.txt
index 3e960b4c77a531221739831ef70e2348e407fd1c..31875348b6f55bb5513dd3785c91d0cddb1a671e 100644
--- a/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.Block.rst.txt
+++ b/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.Block.rst.txt
@@ -1,4 +1,4 @@
-pysegcnn.core.layers.Block
+pysegcnn.core.layers.Block
==========================
.. currentmodule:: pysegcnn.core.layers
diff --git a/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.Conv2dSame.rst.txt b/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.Conv2dSame.rst.txt
index 7c81180f3dbf6d1b08ea77e608a908975395ddb5..42b472e1ec390a7329cfc9813c4a19efb35fe0ff 100644
--- a/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.Conv2dSame.rst.txt
+++ b/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.Conv2dSame.rst.txt
@@ -1,4 +1,4 @@
-pysegcnn.core.layers.Conv2dSame
+pysegcnn.core.layers.Conv2dSame
===============================
.. currentmodule:: pysegcnn.core.layers
diff --git a/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.ConvBnReluMaxPool.rst.txt b/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.ConvBnReluMaxPool.rst.txt
index 7a99902ee622930c6a3ed940f8f2f7e8af705b9d..6260fe50769352e915c3dc3d26ae9b9310ea7c52 100644
--- a/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.ConvBnReluMaxPool.rst.txt
+++ b/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.ConvBnReluMaxPool.rst.txt
@@ -1,4 +1,4 @@
-pysegcnn.core.layers.ConvBnReluMaxPool
+pysegcnn.core.layers.ConvBnReluMaxPool
======================================
.. currentmodule:: pysegcnn.core.layers
diff --git a/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.ConvBnReluMaxUnpool.rst.txt b/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.ConvBnReluMaxUnpool.rst.txt
index c83252a532e4d66427284193e0ce3bd6233ad52e..386abc1be48bcf7487098b4847c7e6385d849772 100644
--- a/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.ConvBnReluMaxUnpool.rst.txt
+++ b/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.ConvBnReluMaxUnpool.rst.txt
@@ -1,4 +1,4 @@
-pysegcnn.core.layers.ConvBnReluMaxUnpool
+pysegcnn.core.layers.ConvBnReluMaxUnpool
========================================
.. currentmodule:: pysegcnn.core.layers
diff --git a/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.ConvBnReluUpsample.rst.txt b/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.ConvBnReluUpsample.rst.txt
index bac34eecc38456d240d6889d01a7e97b9268dd99..6d4f3e6ac7d591f5aaa1da607b6c5730d099d6d0 100644
--- a/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.ConvBnReluUpsample.rst.txt
+++ b/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.ConvBnReluUpsample.rst.txt
@@ -1,4 +1,4 @@
-pysegcnn.core.layers.ConvBnReluUpsample
+pysegcnn.core.layers.ConvBnReluUpsample
=======================================
.. currentmodule:: pysegcnn.core.layers
diff --git a/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.Decoder.rst.txt b/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.Decoder.rst.txt
index 9ca6527d10479cdf03eb3d4213af41da9c295b4a..a5fcf7fd1a0190303deaee9d603008990cc060f5 100644
--- a/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.Decoder.rst.txt
+++ b/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.Decoder.rst.txt
@@ -1,4 +1,4 @@
-pysegcnn.core.layers.Decoder
+pysegcnn.core.layers.Decoder
============================
.. currentmodule:: pysegcnn.core.layers
diff --git a/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.DecoderBlock.rst.txt b/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.DecoderBlock.rst.txt
index 3acc653c6098b27269e7e171bdc7a9f22ab570e8..d6a87bc7f63568ac82cddf0dc5f4d4653391343c 100644
--- a/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.DecoderBlock.rst.txt
+++ b/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.DecoderBlock.rst.txt
@@ -1,4 +1,4 @@
-pysegcnn.core.layers.DecoderBlock
+pysegcnn.core.layers.DecoderBlock
=================================
.. currentmodule:: pysegcnn.core.layers
diff --git a/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.Encoder.rst.txt b/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.Encoder.rst.txt
index 09aa71ccc1c2f5cf0c5fcf0b84a6e47fde490f4e..fcd00c6bc912c46d3ac641b3d2fdeb5a1fd3295c 100644
--- a/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.Encoder.rst.txt
+++ b/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.Encoder.rst.txt
@@ -1,4 +1,4 @@
-pysegcnn.core.layers.Encoder
+pysegcnn.core.layers.Encoder
============================
.. currentmodule:: pysegcnn.core.layers
diff --git a/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.EncoderBlock.rst.txt b/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.EncoderBlock.rst.txt
index 37385b204208ec53e9a247657998a864ad5a0602..454ce4102139d897e53bbad4c61d61bc7b6aeae6 100644
--- a/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.EncoderBlock.rst.txt
+++ b/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.EncoderBlock.rst.txt
@@ -1,4 +1,4 @@
-pysegcnn.core.layers.EncoderBlock
+pysegcnn.core.layers.EncoderBlock
=================================
.. currentmodule:: pysegcnn.core.layers
diff --git a/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.rst.txt b/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.rst.txt
deleted file mode 100644
index 5e2f12505e569b7122b7c49d457efd349481ebfa..0000000000000000000000000000000000000000
--- a/docs/_build/html/_sources/source/generated/pysegcnn.core.layers.rst.txt
+++ /dev/null
@@ -1,43 +0,0 @@
-pysegcnn.core.layers
-====================
-
-.. automodule:: pysegcnn.core.layers
-
-
-
-
-
-
-
- .. rubric:: Functions
-
- .. autosummary::
-
- conv_bn_relu
-
-
-
-
-
- .. rubric:: Classes
-
- .. autosummary::
-
- Block
- Conv2dSame
- ConvBnReluMaxPool
- ConvBnReluMaxUnpool
- ConvBnReluUpsample
- Decoder
- DecoderBlock
- Encoder
- EncoderBlock
-
-
-
-
-
-
-
-
-
diff --git a/docs/_build/html/_sources/source/generated/pysegcnn.core.models.Network.rst.txt b/docs/_build/html/_sources/source/generated/pysegcnn.core.models.Network.rst.txt
index a5b21f07fee5ee799539ac7daf52a68abd48d617..99819a3608b928852226184cbf9d39f50697a252 100644
--- a/docs/_build/html/_sources/source/generated/pysegcnn.core.models.Network.rst.txt
+++ b/docs/_build/html/_sources/source/generated/pysegcnn.core.models.Network.rst.txt
@@ -1,4 +1,4 @@
-pysegcnn.core.models.Network
+pysegcnn.core.models.Network
============================
.. currentmodule:: pysegcnn.core.models
diff --git a/docs/_build/html/_sources/source/generated/pysegcnn.core.models.UNet.rst.txt b/docs/_build/html/_sources/source/generated/pysegcnn.core.models.UNet.rst.txt
index 933d3fcbccf6d0d26a1c0bee26300e0fa0cae620..92cb118b24485a58a9d65aafbd223aafadf1ce59 100644
--- a/docs/_build/html/_sources/source/generated/pysegcnn.core.models.UNet.rst.txt
+++ b/docs/_build/html/_sources/source/generated/pysegcnn.core.models.UNet.rst.txt
@@ -1,4 +1,4 @@
-pysegcnn.core.models.UNet
+pysegcnn.core.models.UNet
=========================
.. currentmodule:: pysegcnn.core.models
diff --git a/docs/_build/html/_sources/source/generated/pysegcnn.core.models.rst.txt b/docs/_build/html/_sources/source/generated/pysegcnn.core.models.rst.txt
deleted file mode 100644
index 3c9f372176cd6078f5561b4a40a331577f0593b9..0000000000000000000000000000000000000000
--- a/docs/_build/html/_sources/source/generated/pysegcnn.core.models.rst.txt
+++ /dev/null
@@ -1,33 +0,0 @@
-pysegcnn.core.models
-====================
-
-.. automodule:: pysegcnn.core.models
-
-
-
-
-
-
-
-
-
-
-
- .. rubric:: Classes
-
- .. autosummary::
-
- Network
- SupportedLossFunctions
- SupportedModels
- SupportedOptimizers
- UNet
-
-
-
-
-
-
-
-
-
diff --git a/docs/_build/html/_sources/source/install.rst.txt b/docs/_build/html/_sources/source/install.rst.txt
deleted file mode 100644
index eb91803d2c6ba64ea68e39a298feb2eaf9ea1254..0000000000000000000000000000000000000000
--- a/docs/_build/html/_sources/source/install.rst.txt
+++ /dev/null
@@ -1,77 +0,0 @@
-Installation
-============
-
-Download
----------
-You can download ``pysegcnn`` from this repository's
-`website <https://gitlab.inf.unibz.it/REMSEN/ccisnow/pysegcnn>`_
-or alternatively use ``git`` from terminal:
-
-.. code-block:: bash
-
- git clone https://gitlab.inf.unibz.it/REMSEN/ccisnow/pysegcnn.git
-
-This creates a copy of the repository in your current directory on the file
-system.
-
-Conda
------
-
-To install ``pysegcnn``, I recommend to use the ``conda`` package manager.
-You can download ``conda`` `here <https://docs.conda.io/en/latest/miniconda.html>`_.
-Once successfully installed ``conda``, I recommend to add ``conda-forge`` as
-your default channel:
-
-.. code-block:: bash
-
- conda config --add channels conda-forge
-
-Conda environments
-------------------
-
-To install ``pysegcnn``, I recommend to create a specific ``conda``
-`environment <https://docs.conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html>`_,
-by using the provided ``environment.yml`` file. In a terminal, navigate to the
-**cloned git repositories root directory** (``/pysegcnn``) and type:
-
-.. code-block:: bash
-
- conda env create -f environment.yml
-
-This may take a while. The above command creates a conda environment with all
-required dependencies except the ``pytorch`` package. The first line in
-``environment.yml`` defines the environment name, in this case ``pysegcnn``.
-Activate your environment using:
-
-.. code-block:: bash
-
- conda activate pysegcnn
-
-Install PyTorch
----------------
-The installation of ``pytorch`` is heavily dependent on the hardware of your
-machine. Therefore, after activating your environment, install the version of
-`PyTorch <https://pytorch.org/>`_ that your system supports by following the
-official `instructions <https://pytorch.org/get-started/locally/>`_.
-
-If you have to build ``pytorch`` from source, follow this
-`guide <https://github.com/pytorch/pytorch#from-source>`_.
-
-`Here <https://pytorch.org/get-started/locally/#linux-verification>`_ is how to
-check if ``pytorch`` is installed correctly.
-
-Install PySegCNN
-----------------
-To finally install ``pysegcnn`` run the below command **from this repositories
-root directory within the activated ``pysegcnn`` conda environment**:
-
-.. code-block:: bash
-
- pip install -e .
-
-If successful, you should be able to import ``pysegcnn`` from any Python
-interpreter using:
-
-.. code-block:: python
-
- import pysegcnn
diff --git a/docs/_build/html/_sources/source/main.rst.txt b/docs/_build/html/_sources/source/main.rst.txt
deleted file mode 100644
index 8fd37169d80e63c1a1e29b5dba6e2f67eb27f43b..0000000000000000000000000000000000000000
--- a/docs/_build/html/_sources/source/main.rst.txt
+++ /dev/null
@@ -1,37 +0,0 @@
-main package
-============
-
-Submodules
-----------
-
-main.config module
-------------------
-
-.. automodule:: main.config
- :members:
- :undoc-members:
- :show-inheritance:
-
-main.eval module
-----------------
-
-.. automodule:: main.eval
- :members:
- :undoc-members:
- :show-inheritance:
-
-main.train module
------------------
-
-.. automodule:: main.train
- :members:
- :undoc-members:
- :show-inheritance:
-
-Module contents
----------------
-
-.. automodule:: main
- :members:
- :undoc-members:
- :show-inheritance:
diff --git a/docs/_build/html/_sources/source/modules.rst.txt b/docs/_build/html/_sources/source/modules.rst.txt
deleted file mode 100644
index 55e86457961793c751b95f0118225c88718c3100..0000000000000000000000000000000000000000
--- a/docs/_build/html/_sources/source/modules.rst.txt
+++ /dev/null
@@ -1,9 +0,0 @@
-pysegcnn
-========
-
-.. toctree::
- :maxdepth: 4
-
- core
- main
- preprocessing
diff --git a/docs/_build/html/_sources/source/preprocessing.rst.txt b/docs/_build/html/_sources/source/preprocessing.rst.txt
deleted file mode 100644
index 06afc1c0ea6732e982af1a848fcd9baef6a8503e..0000000000000000000000000000000000000000
--- a/docs/_build/html/_sources/source/preprocessing.rst.txt
+++ /dev/null
@@ -1,21 +0,0 @@
-preprocessing package
-=====================
-
-Submodules
-----------
-
-preprocessing.sparcs module
----------------------------
-
-.. automodule:: preprocessing.sparcs
- :members:
- :undoc-members:
- :show-inheritance:
-
-Module contents
----------------
-
-.. automodule:: preprocessing
- :members:
- :undoc-members:
- :show-inheritance:
diff --git a/docs/_build/html/_sources/source/readme.rst.txt b/docs/_build/html/_sources/source/readme.rst.txt
deleted file mode 100644
index 3f2eb6fc98784afb1aa1a507ef51c1f457ca7689..0000000000000000000000000000000000000000
--- a/docs/_build/html/_sources/source/readme.rst.txt
+++ /dev/null
@@ -1,4 +0,0 @@
-README
-======
-
-.. include: ../../README.md
\ No newline at end of file
diff --git a/docs/_build/html/_static/alabaster.css b/docs/_build/html/_static/alabaster.css
deleted file mode 100644
index 0eddaeb07d19bffb50884d7bd7996418b6461a09..0000000000000000000000000000000000000000
--- a/docs/_build/html/_static/alabaster.css
+++ /dev/null
@@ -1,701 +0,0 @@
-@import url("basic.css");
-
-/* -- page layout ----------------------------------------------------------- */
-
-body {
- font-family: Georgia, serif;
- font-size: 17px;
- background-color: #fff;
- color: #000;
- margin: 0;
- padding: 0;
-}
-
-
-div.document {
- width: 940px;
- margin: 30px auto 0 auto;
-}
-
-div.documentwrapper {
- float: left;
- width: 100%;
-}
-
-div.bodywrapper {
- margin: 0 0 0 220px;
-}
-
-div.sphinxsidebar {
- width: 220px;
- font-size: 14px;
- line-height: 1.5;
-}
-
-hr {
- border: 1px solid #B1B4B6;
-}
-
-div.body {
- background-color: #fff;
- color: #3E4349;
- padding: 0 30px 0 30px;
-}
-
-div.body > .section {
- text-align: left;
-}
-
-div.footer {
- width: 940px;
- margin: 20px auto 30px auto;
- font-size: 14px;
- color: #888;
- text-align: right;
-}
-
-div.footer a {
- color: #888;
-}
-
-p.caption {
- font-family: inherit;
- font-size: inherit;
-}
-
-
-div.relations {
- display: none;
-}
-
-
-div.sphinxsidebar a {
- color: #444;
- text-decoration: none;
- border-bottom: 1px dotted #999;
-}
-
-div.sphinxsidebar a:hover {
- border-bottom: 1px solid #999;
-}
-
-div.sphinxsidebarwrapper {
- padding: 18px 10px;
-}
-
-div.sphinxsidebarwrapper p.logo {
- padding: 0;
- margin: -10px 0 0 0px;
- text-align: center;
-}
-
-div.sphinxsidebarwrapper h1.logo {
- margin-top: -10px;
- text-align: center;
- margin-bottom: 5px;
- text-align: left;
-}
-
-div.sphinxsidebarwrapper h1.logo-name {
- margin-top: 0px;
-}
-
-div.sphinxsidebarwrapper p.blurb {
- margin-top: 0;
- font-style: normal;
-}
-
-div.sphinxsidebar h3,
-div.sphinxsidebar h4 {
- font-family: Georgia, serif;
- color: #444;
- font-size: 24px;
- font-weight: normal;
- margin: 0 0 5px 0;
- padding: 0;
-}
-
-div.sphinxsidebar h4 {
- font-size: 20px;
-}
-
-div.sphinxsidebar h3 a {
- color: #444;
-}
-
-div.sphinxsidebar p.logo a,
-div.sphinxsidebar h3 a,
-div.sphinxsidebar p.logo a:hover,
-div.sphinxsidebar h3 a:hover {
- border: none;
-}
-
-div.sphinxsidebar p {
- color: #555;
- margin: 10px 0;
-}
-
-div.sphinxsidebar ul {
- margin: 10px 0;
- padding: 0;
- color: #000;
-}
-
-div.sphinxsidebar ul li.toctree-l1 > a {
- font-size: 120%;
-}
-
-div.sphinxsidebar ul li.toctree-l2 > a {
- font-size: 110%;
-}
-
-div.sphinxsidebar input {
- border: 1px solid #CCC;
- font-family: Georgia, serif;
- font-size: 1em;
-}
-
-div.sphinxsidebar hr {
- border: none;
- height: 1px;
- color: #AAA;
- background: #AAA;
-
- text-align: left;
- margin-left: 0;
- width: 50%;
-}
-
-div.sphinxsidebar .badge {
- border-bottom: none;
-}
-
-div.sphinxsidebar .badge:hover {
- border-bottom: none;
-}
-
-/* To address an issue with donation coming after search */
-div.sphinxsidebar h3.donation {
- margin-top: 10px;
-}
-
-/* -- body styles ----------------------------------------------------------- */
-
-a {
- color: #004B6B;
- text-decoration: underline;
-}
-
-a:hover {
- color: #6D4100;
- text-decoration: underline;
-}
-
-div.body h1,
-div.body h2,
-div.body h3,
-div.body h4,
-div.body h5,
-div.body h6 {
- font-family: Georgia, serif;
- font-weight: normal;
- margin: 30px 0px 10px 0px;
- padding: 0;
-}
-
-div.body h1 { margin-top: 0; padding-top: 0; font-size: 240%; }
-div.body h2 { font-size: 180%; }
-div.body h3 { font-size: 150%; }
-div.body h4 { font-size: 130%; }
-div.body h5 { font-size: 100%; }
-div.body h6 { font-size: 100%; }
-
-a.headerlink {
- color: #DDD;
- padding: 0 4px;
- text-decoration: none;
-}
-
-a.headerlink:hover {
- color: #444;
- background: #EAEAEA;
-}
-
-div.body p, div.body dd, div.body li {
- line-height: 1.4em;
-}
-
-div.admonition {
- margin: 20px 0px;
- padding: 10px 30px;
- background-color: #EEE;
- border: 1px solid #CCC;
-}
-
-div.admonition tt.xref, div.admonition code.xref, div.admonition a tt {
- background-color: #FBFBFB;
- border-bottom: 1px solid #fafafa;
-}
-
-div.admonition p.admonition-title {
- font-family: Georgia, serif;
- font-weight: normal;
- font-size: 24px;
- margin: 0 0 10px 0;
- padding: 0;
- line-height: 1;
-}
-
-div.admonition p.last {
- margin-bottom: 0;
-}
-
-div.highlight {
- background-color: #fff;
-}
-
-dt:target, .highlight {
- background: #FAF3E8;
-}
-
-div.warning {
- background-color: #FCC;
- border: 1px solid #FAA;
-}
-
-div.danger {
- background-color: #FCC;
- border: 1px solid #FAA;
- -moz-box-shadow: 2px 2px 4px #D52C2C;
- -webkit-box-shadow: 2px 2px 4px #D52C2C;
- box-shadow: 2px 2px 4px #D52C2C;
-}
-
-div.error {
- background-color: #FCC;
- border: 1px solid #FAA;
- -moz-box-shadow: 2px 2px 4px #D52C2C;
- -webkit-box-shadow: 2px 2px 4px #D52C2C;
- box-shadow: 2px 2px 4px #D52C2C;
-}
-
-div.caution {
- background-color: #FCC;
- border: 1px solid #FAA;
-}
-
-div.attention {
- background-color: #FCC;
- border: 1px solid #FAA;
-}
-
-div.important {
- background-color: #EEE;
- border: 1px solid #CCC;
-}
-
-div.note {
- background-color: #EEE;
- border: 1px solid #CCC;
-}
-
-div.tip {
- background-color: #EEE;
- border: 1px solid #CCC;
-}
-
-div.hint {
- background-color: #EEE;
- border: 1px solid #CCC;
-}
-
-div.seealso {
- background-color: #EEE;
- border: 1px solid #CCC;
-}
-
-div.topic {
- background-color: #EEE;
-}
-
-p.admonition-title {
- display: inline;
-}
-
-p.admonition-title:after {
- content: ":";
-}
-
-pre, tt, code {
- font-family: 'Consolas', 'Menlo', 'DejaVu Sans Mono', 'Bitstream Vera Sans Mono', monospace;
- font-size: 0.9em;
-}
-
-.hll {
- background-color: #FFC;
- margin: 0 -12px;
- padding: 0 12px;
- display: block;
-}
-
-img.screenshot {
-}
-
-tt.descname, tt.descclassname, code.descname, code.descclassname {
- font-size: 0.95em;
-}
-
-tt.descname, code.descname {
- padding-right: 0.08em;
-}
-
-img.screenshot {
- -moz-box-shadow: 2px 2px 4px #EEE;
- -webkit-box-shadow: 2px 2px 4px #EEE;
- box-shadow: 2px 2px 4px #EEE;
-}
-
-table.docutils {
- border: 1px solid #888;
- -moz-box-shadow: 2px 2px 4px #EEE;
- -webkit-box-shadow: 2px 2px 4px #EEE;
- box-shadow: 2px 2px 4px #EEE;
-}
-
-table.docutils td, table.docutils th {
- border: 1px solid #888;
- padding: 0.25em 0.7em;
-}
-
-table.field-list, table.footnote {
- border: none;
- -moz-box-shadow: none;
- -webkit-box-shadow: none;
- box-shadow: none;
-}
-
-table.footnote {
- margin: 15px 0;
- width: 100%;
- border: 1px solid #EEE;
- background: #FDFDFD;
- font-size: 0.9em;
-}
-
-table.footnote + table.footnote {
- margin-top: -15px;
- border-top: none;
-}
-
-table.field-list th {
- padding: 0 0.8em 0 0;
-}
-
-table.field-list td {
- padding: 0;
-}
-
-table.field-list p {
- margin-bottom: 0.8em;
-}
-
-/* Cloned from
- * https://github.com/sphinx-doc/sphinx/commit/ef60dbfce09286b20b7385333d63a60321784e68
- */
-.field-name {
- -moz-hyphens: manual;
- -ms-hyphens: manual;
- -webkit-hyphens: manual;
- hyphens: manual;
-}
-
-table.footnote td.label {
- width: .1px;
- padding: 0.3em 0 0.3em 0.5em;
-}
-
-table.footnote td {
- padding: 0.3em 0.5em;
-}
-
-dl {
- margin: 0;
- padding: 0;
-}
-
-dl dd {
- margin-left: 30px;
-}
-
-blockquote {
- margin: 0 0 0 30px;
- padding: 0;
-}
-
-ul, ol {
- /* Matches the 30px from the narrow-screen "li > ul" selector below */
- margin: 10px 0 10px 30px;
- padding: 0;
-}
-
-pre {
- background: #EEE;
- padding: 7px 30px;
- margin: 15px 0px;
- line-height: 1.3em;
-}
-
-div.viewcode-block:target {
- background: #ffd;
-}
-
-dl pre, blockquote pre, li pre {
- margin-left: 0;
- padding-left: 30px;
-}
-
-tt, code {
- background-color: #ecf0f3;
- color: #222;
- /* padding: 1px 2px; */
-}
-
-tt.xref, code.xref, a tt {
- background-color: #FBFBFB;
- border-bottom: 1px solid #fff;
-}
-
-a.reference {
- text-decoration: none;
- border-bottom: 1px dotted #004B6B;
-}
-
-/* Don't put an underline on images */
-a.image-reference, a.image-reference:hover {
- border-bottom: none;
-}
-
-a.reference:hover {
- border-bottom: 1px solid #6D4100;
-}
-
-a.footnote-reference {
- text-decoration: none;
- font-size: 0.7em;
- vertical-align: top;
- border-bottom: 1px dotted #004B6B;
-}
-
-a.footnote-reference:hover {
- border-bottom: 1px solid #6D4100;
-}
-
-a:hover tt, a:hover code {
- background: #EEE;
-}
-
-
-@media screen and (max-width: 870px) {
-
- div.sphinxsidebar {
- display: none;
- }
-
- div.document {
- width: 100%;
-
- }
-
- div.documentwrapper {
- margin-left: 0;
- margin-top: 0;
- margin-right: 0;
- margin-bottom: 0;
- }
-
- div.bodywrapper {
- margin-top: 0;
- margin-right: 0;
- margin-bottom: 0;
- margin-left: 0;
- }
-
- ul {
- margin-left: 0;
- }
-
- li > ul {
- /* Matches the 30px from the "ul, ol" selector above */
- margin-left: 30px;
- }
-
- .document {
- width: auto;
- }
-
- .footer {
- width: auto;
- }
-
- .bodywrapper {
- margin: 0;
- }
-
- .footer {
- width: auto;
- }
-
- .github {
- display: none;
- }
-
-
-
-}
-
-
-
-@media screen and (max-width: 875px) {
-
- body {
- margin: 0;
- padding: 20px 30px;
- }
-
- div.documentwrapper {
- float: none;
- background: #fff;
- }
-
- div.sphinxsidebar {
- display: block;
- float: none;
- width: 102.5%;
- margin: 50px -30px -20px -30px;
- padding: 10px 20px;
- background: #333;
- color: #FFF;
- }
-
- div.sphinxsidebar h3, div.sphinxsidebar h4, div.sphinxsidebar p,
- div.sphinxsidebar h3 a {
- color: #fff;
- }
-
- div.sphinxsidebar a {
- color: #AAA;
- }
-
- div.sphinxsidebar p.logo {
- display: none;
- }
-
- div.document {
- width: 100%;
- margin: 0;
- }
-
- div.footer {
- display: none;
- }
-
- div.bodywrapper {
- margin: 0;
- }
-
- div.body {
- min-height: 0;
- padding: 0;
- }
-
- .rtd_doc_footer {
- display: none;
- }
-
- .document {
- width: auto;
- }
-
- .footer {
- width: auto;
- }
-
- .footer {
- width: auto;
- }
-
- .github {
- display: none;
- }
-}
-
-
-/* misc. */
-
-.revsys-inline {
- display: none!important;
-}
-
-/* Make nested-list/multi-paragraph items look better in Releases changelog
- * pages. Without this, docutils' magical list fuckery causes inconsistent
- * formatting between different release sub-lists.
- */
-div#changelog > div.section > ul > li > p:only-child {
- margin-bottom: 0;
-}
-
-/* Hide fugly table cell borders in ..bibliography:: directive output */
-table.docutils.citation, table.docutils.citation td, table.docutils.citation th {
- border: none;
- /* Below needed in some edge cases; if not applied, bottom shadows appear */
- -moz-box-shadow: none;
- -webkit-box-shadow: none;
- box-shadow: none;
-}
-
-
-/* relbar */
-
-.related {
- line-height: 30px;
- width: 100%;
- font-size: 0.9rem;
-}
-
-.related.top {
- border-bottom: 1px solid #EEE;
- margin-bottom: 20px;
-}
-
-.related.bottom {
- border-top: 1px solid #EEE;
-}
-
-.related ul {
- padding: 0;
- margin: 0;
- list-style: none;
-}
-
-.related li {
- display: inline;
-}
-
-nav#rellinks {
- float: right;
-}
-
-nav#rellinks li+li:before {
- content: "|";
-}
-
-nav#breadcrumbs li+li:before {
- content: "\00BB";
-}
-
-/* Hide certain items when printing */
-@media print {
- div.related {
- display: none;
- }
-}
\ No newline at end of file
diff --git a/docs/_build/html/_static/custom.css b/docs/_build/html/_static/custom.css
deleted file mode 100644
index 2a924f1d6a8bc930c5296bdb2d5c2d3e39b04a1c..0000000000000000000000000000000000000000
--- a/docs/_build/html/_static/custom.css
+++ /dev/null
@@ -1 +0,0 @@
-/* This file intentionally left blank. */
diff --git a/docs/_build/html/genindex.html b/docs/_build/html/genindex.html
index 7fb412d23cd3e8c8871c22f2206c79156b5bc78b..d44bdbf056114006563185ef8fe370b6fb665423 100644
--- a/docs/_build/html/genindex.html
+++ b/docs/_build/html/genindex.html
@@ -152,6 +152,7 @@
<div class="genindex-jumpbox">
<a href="#_"><strong>_</strong></a>
+ | <a href="#A"><strong>A</strong></a>
| <a href="#B"><strong>B</strong></a>
| <a href="#C"><strong>C</strong></a>
| <a href="#D"><strong>D</strong></a>
@@ -171,6 +172,8 @@
<li><a href="source/generated/pysegcnn.core.dataset.ImageDataset.html#pysegcnn.core.dataset.ImageDataset.__init__">(pysegcnn.core.dataset.ImageDataset method)</a>
</li>
<li><a href="source/generated/pysegcnn.core.dataset.SparcsDataset.html#pysegcnn.core.dataset.SparcsDataset.__init__">(pysegcnn.core.dataset.SparcsDataset method)</a>
+</li>
+ <li><a href="source/generated/pysegcnn.core.dataset.StandardEoDataset.html#pysegcnn.core.dataset.StandardEoDataset.__init__">(pysegcnn.core.dataset.StandardEoDataset method)</a>
</li>
<li><a href="source/generated/pysegcnn.core.layers.Block.html#pysegcnn.core.layers.Block.__init__">(pysegcnn.core.layers.Block method)</a>
</li>
@@ -193,11 +196,21 @@
<li><a href="source/generated/pysegcnn.core.models.Network.html#pysegcnn.core.models.Network.__init__">(pysegcnn.core.models.Network method)</a>
</li>
<li><a href="source/generated/pysegcnn.core.models.UNet.html#pysegcnn.core.models.UNet.__init__">(pysegcnn.core.models.UNet method)</a>
+</li>
+ <li><a href="source/generated/pysegcnn.core.transforms.Augment.html#pysegcnn.core.transforms.Augment.__init__">(pysegcnn.core.transforms.Augment method)</a>
</li>
</ul></li>
</ul></td>
</tr></table>
+<h2 id="A">A</h2>
+<table style="width: 100%" class="indextable genindextable"><tr>
+ <td style="width: 33%; vertical-align: top;"><ul>
+ <li><a href="source/generated/pysegcnn.core.transforms.Augment.html#pysegcnn.core.transforms.Augment">Augment (class in pysegcnn.core.transforms)</a>
+</li>
+ </ul></td>
+</tr></table>
+
<h2 id="B">B</h2>
<table style="width: 100%" class="indextable genindextable"><tr>
<td style="width: 33%; vertical-align: top;"><ul>
@@ -268,6 +281,10 @@
<table style="width: 100%" class="indextable genindextable"><tr>
<td style="width: 33%; vertical-align: top;"><ul>
<li><a href="source/generated/pysegcnn.core.dataset.SparcsDataset.html#pysegcnn.core.dataset.SparcsDataset">SparcsDataset (class in pysegcnn.core.dataset)</a>
+</li>
+ </ul></td>
+ <td style="width: 33%; vertical-align: top;"><ul>
+ <li><a href="source/generated/pysegcnn.core.dataset.StandardEoDataset.html#pysegcnn.core.dataset.StandardEoDataset">StandardEoDataset (class in pysegcnn.core.dataset)</a>
</li>
</ul></td>
</tr></table>
diff --git a/docs/_build/html/modules.html b/docs/_build/html/modules.html
deleted file mode 100644
index a1bf15a47e95fddfeda69bf11cd5d32742020290..0000000000000000000000000000000000000000
--- a/docs/_build/html/modules.html
+++ /dev/null
@@ -1,131 +0,0 @@
-
-<!DOCTYPE html>
-
-<html>
- <head>
- <meta charset="utf-8" />
- <meta name="viewport" content="width=device-width, initial-scale=1.0" />
- <title>Modules — PySegCNN documentation</title>
- <link rel="stylesheet" href="_static/alabaster.css" type="text/css" />
- <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
- <script id="documentation_options" data-url_root="./" src="_static/documentation_options.js"></script>
- <script src="_static/jquery.js"></script>
- <script src="_static/underscore.js"></script>
- <script src="_static/doctools.js"></script>
- <script src="_static/language_data.js"></script>
- <link rel="index" title="Index" href="genindex.html" />
- <link rel="search" title="Search" href="search.html" />
- <link rel="next" title="pysegcnn.core.cli" href="modules/pysegcnn.core.cli.html" />
- <link rel="prev" title="PySegCNN: Image segmentation with convolutional neural networks in Python" href="index.html" />
-
- <link rel="stylesheet" href="_static/custom.css" type="text/css" />
-
-
- <meta name="viewport" content="width=device-width, initial-scale=0.9, maximum-scale=0.9" />
-
- </head><body>
-
-
- <div class="document">
- <div class="documentwrapper">
- <div class="bodywrapper">
-
-
- <div class="body" role="main">
-
- <div class="section" id="modules">
-<h1>Modules<a class="headerlink" href="#modules" title="Permalink to this headline">¶</a></h1>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><a class="reference internal" href="modules/pysegcnn.core.cli.html#module-pysegcnn.core.cli" title="pysegcnn.core.cli"><code class="xref py py-obj docutils literal notranslate"><span class="pre">pysegcnn.core.cli</span></code></a></p></td>
-<td><p>Command line interface parsers.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="modules/pysegcnn.core.constants.html#module-pysegcnn.core.constants" title="pysegcnn.core.constants"><code class="xref py py-obj docutils literal notranslate"><span class="pre">pysegcnn.core.constants</span></code></a></p></td>
-<td><p>A collection of constant values.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="modules/pysegcnn.core.dataset.html#module-pysegcnn.core.dataset" title="pysegcnn.core.dataset"><code class="xref py py-obj docutils literal notranslate"><span class="pre">pysegcnn.core.dataset</span></code></a></p></td>
-<td><p>Custom dataset classes compliant to the PyTorch standard.</p></td>
-</tr>
-</tbody>
-</table>
-</div>
-
-
- </div>
-
- </div>
- </div>
- <div class="sphinxsidebar" role="navigation" aria-label="main navigation">
- <div class="sphinxsidebarwrapper">
-<h1 class="logo"><a href="index.html">PySegCNN</a></h1>
-
-
-
-
-
-
-
-
-<h3>Navigation</h3>
-<p class="caption"><span class="caption-text">Table of contents:</span></p>
-<ul class="current">
-<li class="toctree-l1 current"><a class="current reference internal" href="#">Modules</a><ul>
-<li class="toctree-l2"><a class="reference internal" href="modules/pysegcnn.core.cli.html">pysegcnn.core.cli</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/pysegcnn.core.constants.html">pysegcnn.core.constants</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/pysegcnn.core.dataset.html">pysegcnn.core.dataset</a></li>
-</ul>
-</li>
-</ul>
-
-<div class="relations">
-<h3>Related Topics</h3>
-<ul>
- <li><a href="index.html">Documentation overview</a><ul>
- <li>Previous: <a href="index.html" title="previous chapter">PySegCNN: Image segmentation with convolutional neural networks in Python</a></li>
- <li>Next: <a href="modules/pysegcnn.core.cli.html" title="next chapter">pysegcnn.core.cli</a></li>
- </ul></li>
-</ul>
-</div>
-<div id="searchbox" style="display: none" role="search">
- <h3 id="searchlabel">Quick search</h3>
- <div class="searchformwrapper">
- <form class="search" action="search.html" method="get">
- <input type="text" name="q" aria-labelledby="searchlabel" />
- <input type="submit" value="Go" />
- </form>
- </div>
-</div>
-<script>$('#searchbox').show(0);</script>
-
-
-
-
-
-
-
-
- </div>
- </div>
- <div class="clearer"></div>
- </div>
- <div class="footer">
- ©2020, Daniel Frisinghelli.
-
- |
- Powered by <a href="http://sphinx-doc.org/">Sphinx 3.2.1</a>
- & <a href="https://github.com/bitprophet/alabaster">Alabaster 0.7.12</a>
-
- |
- <a href="_sources/modules.rst.txt"
- rel="nofollow">Page source</a>
- </div>
-
-
-
-
- </body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/modules/pysegcnn.core.cli.html b/docs/_build/html/modules/pysegcnn.core.cli.html
deleted file mode 100644
index 9c3e68b4583f7cb1bbb0253ee557bbe44a5b268d..0000000000000000000000000000000000000000
--- a/docs/_build/html/modules/pysegcnn.core.cli.html
+++ /dev/null
@@ -1,137 +0,0 @@
-
-<!DOCTYPE html>
-
-<html>
- <head>
- <meta charset="utf-8" />
- <meta name="viewport" content="width=device-width, initial-scale=1.0" />
- <title>pysegcnn.core.cli — PySegCNN documentation</title>
- <link rel="stylesheet" href="../_static/alabaster.css" type="text/css" />
- <link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
- <script id="documentation_options" data-url_root="../" src="../_static/documentation_options.js"></script>
- <script src="../_static/jquery.js"></script>
- <script src="../_static/underscore.js"></script>
- <script src="../_static/doctools.js"></script>
- <script src="../_static/language_data.js"></script>
- <link rel="index" title="Index" href="../genindex.html" />
- <link rel="search" title="Search" href="../search.html" />
- <link rel="next" title="pysegcnn.core.constants" href="pysegcnn.core.constants.html" />
- <link rel="prev" title="Modules" href="../modules.html" />
-
- <link rel="stylesheet" href="../_static/custom.css" type="text/css" />
-
-
- <meta name="viewport" content="width=device-width, initial-scale=0.9, maximum-scale=0.9" />
-
- </head><body>
-
-
- <div class="document">
- <div class="documentwrapper">
- <div class="bodywrapper">
-
-
- <div class="body" role="main">
-
- <div class="section" id="module-pysegcnn.core.cli">
-<span id="pysegcnn-core-cli"></span><h1>pysegcnn.core.cli<a class="headerlink" href="#module-pysegcnn.core.cli" title="Permalink to this headline">¶</a></h1>
-<p>Command line interface parsers.</p>
-<div class="section" id="license">
-<h2>License<a class="headerlink" href="#license" title="Permalink to this headline">¶</a></h2>
-<blockquote>
-<div><p>Copyright (c) 2020 Daniel Frisinghelli</p>
-<p>This source code is licensed under the GNU General Public License v3.</p>
-<p>See the LICENSE file in the repository’s root directory.</p>
-</div></blockquote>
-<p class="rubric">Functions</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">structure_parser</span></code>()</p></td>
-<td><p>Command line argument parser to standardize dataset structure.</p></td>
-</tr>
-</tbody>
-</table>
-</div>
-</div>
-
-
- </div>
-
- </div>
- </div>
- <div class="sphinxsidebar" role="navigation" aria-label="main navigation">
- <div class="sphinxsidebarwrapper">
-<h1 class="logo"><a href="../index.html">PySegCNN</a></h1>
-
-
-
-
-
-
-
-
-<h3>Navigation</h3>
-<p class="caption"><span class="caption-text">Table of contents:</span></p>
-<ul class="current">
-<li class="toctree-l1 current"><a class="reference internal" href="../modules.html">Modules</a><ul class="current">
-<li class="toctree-l2 current"><a class="current reference internal" href="#">pysegcnn.core.cli</a></li>
-<li class="toctree-l2"><a class="reference internal" href="pysegcnn.core.constants.html">pysegcnn.core.constants</a></li>
-<li class="toctree-l2"><a class="reference internal" href="pysegcnn.core.dataset.html">pysegcnn.core.dataset</a></li>
-</ul>
-</li>
-</ul>
-
-<div class="relations">
-<h3>Related Topics</h3>
-<ul>
- <li><a href="../index.html">Documentation overview</a><ul>
- <li><a href="../modules.html">Modules</a><ul>
- <li>Previous: <a href="../modules.html" title="previous chapter">Modules</a></li>
- <li>Next: <a href="pysegcnn.core.constants.html" title="next chapter">pysegcnn.core.constants</a></li>
- </ul></li>
- </ul></li>
-</ul>
-</div>
-<div id="searchbox" style="display: none" role="search">
- <h3 id="searchlabel">Quick search</h3>
- <div class="searchformwrapper">
- <form class="search" action="../search.html" method="get">
- <input type="text" name="q" aria-labelledby="searchlabel" />
- <input type="submit" value="Go" />
- </form>
- </div>
-</div>
-<script>$('#searchbox').show(0);</script>
-
-
-
-
-
-
-
-
- </div>
- </div>
- <div class="clearer"></div>
- </div>
- <div class="footer">
- ©2020, Daniel Frisinghelli.
-
- |
- Powered by <a href="http://sphinx-doc.org/">Sphinx 3.2.1</a>
- & <a href="https://github.com/bitprophet/alabaster">Alabaster 0.7.12</a>
-
- |
- <a href="../_sources/modules/pysegcnn.core.cli.rst.txt"
- rel="nofollow">Page source</a>
- </div>
-
-
-
-
- </body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/modules/pysegcnn.core.constants.html b/docs/_build/html/modules/pysegcnn.core.constants.html
deleted file mode 100644
index 84a61dffd38c1cf0da63b66db2a9ff1467bbb52f..0000000000000000000000000000000000000000
--- a/docs/_build/html/modules/pysegcnn.core.constants.html
+++ /dev/null
@@ -1,152 +0,0 @@
-
-<!DOCTYPE html>
-
-<html>
- <head>
- <meta charset="utf-8" />
- <meta name="viewport" content="width=device-width, initial-scale=1.0" />
- <title>pysegcnn.core.constants — PySegCNN documentation</title>
- <link rel="stylesheet" href="../_static/alabaster.css" type="text/css" />
- <link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
- <script id="documentation_options" data-url_root="../" src="../_static/documentation_options.js"></script>
- <script src="../_static/jquery.js"></script>
- <script src="../_static/underscore.js"></script>
- <script src="../_static/doctools.js"></script>
- <script src="../_static/language_data.js"></script>
- <link rel="index" title="Index" href="../genindex.html" />
- <link rel="search" title="Search" href="../search.html" />
- <link rel="next" title="pysegcnn.core.dataset" href="pysegcnn.core.dataset.html" />
- <link rel="prev" title="pysegcnn.core.cli" href="pysegcnn.core.cli.html" />
-
- <link rel="stylesheet" href="../_static/custom.css" type="text/css" />
-
-
- <meta name="viewport" content="width=device-width, initial-scale=0.9, maximum-scale=0.9" />
-
- </head><body>
-
-
- <div class="document">
- <div class="documentwrapper">
- <div class="bodywrapper">
-
-
- <div class="body" role="main">
-
- <div class="section" id="module-pysegcnn.core.constants">
-<span id="pysegcnn-core-constants"></span><h1>pysegcnn.core.constants<a class="headerlink" href="#module-pysegcnn.core.constants" title="Permalink to this headline">¶</a></h1>
-<p>A collection of constant values.</p>
-<div class="section" id="license">
-<h2>License<a class="headerlink" href="#license" title="Permalink to this headline">¶</a></h2>
-<blockquote>
-<div><p>Copyright (c) 2020 Daniel Frisinghelli</p>
-<p>This source code is licensed under the GNU General Public License v3.</p>
-<p>See the LICENSE file in the repository’s root directory.</p>
-</div></blockquote>
-<p class="rubric">Classes</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">Cloud95Labels</span></code>(value)</p></td>
-<td><p>Class labels of the <a href="#id1"><span class="problematic" id="id2">`Cloud-95`_</span></a> dataset.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">Label</span></code>(value)</p></td>
-<td><p>Generic enumeration for class labels.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">Landsat8</span></code>(value)</p></td>
-<td><p>The spectral bands of the Landsat 8 sensors.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">ProSnowLabels</span></code>(value)</p></td>
-<td><p>Class labels of the ProSnow datasets.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">Sentinel2</span></code>(value)</p></td>
-<td><p>The spectral bands of the Sentinel-2 MultiSpectral Instrument (MSI).</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">SparcsLabels</span></code>(value)</p></td>
-<td><p>Class labels of the <a href="#id3"><span class="problematic" id="id4">`Sparcs`_</span></a> dataset.</p></td>
-</tr>
-</tbody>
-</table>
-</div>
-</div>
-
-
- </div>
-
- </div>
- </div>
- <div class="sphinxsidebar" role="navigation" aria-label="main navigation">
- <div class="sphinxsidebarwrapper">
-<h1 class="logo"><a href="../index.html">PySegCNN</a></h1>
-
-
-
-
-
-
-
-
-<h3>Navigation</h3>
-<p class="caption"><span class="caption-text">Table of contents:</span></p>
-<ul class="current">
-<li class="toctree-l1 current"><a class="reference internal" href="../modules.html">Modules</a><ul class="current">
-<li class="toctree-l2"><a class="reference internal" href="pysegcnn.core.cli.html">pysegcnn.core.cli</a></li>
-<li class="toctree-l2 current"><a class="current reference internal" href="#">pysegcnn.core.constants</a></li>
-<li class="toctree-l2"><a class="reference internal" href="pysegcnn.core.dataset.html">pysegcnn.core.dataset</a></li>
-</ul>
-</li>
-</ul>
-
-<div class="relations">
-<h3>Related Topics</h3>
-<ul>
- <li><a href="../index.html">Documentation overview</a><ul>
- <li><a href="../modules.html">Modules</a><ul>
- <li>Previous: <a href="pysegcnn.core.cli.html" title="previous chapter">pysegcnn.core.cli</a></li>
- <li>Next: <a href="pysegcnn.core.dataset.html" title="next chapter">pysegcnn.core.dataset</a></li>
- </ul></li>
- </ul></li>
-</ul>
-</div>
-<div id="searchbox" style="display: none" role="search">
- <h3 id="searchlabel">Quick search</h3>
- <div class="searchformwrapper">
- <form class="search" action="../search.html" method="get">
- <input type="text" name="q" aria-labelledby="searchlabel" />
- <input type="submit" value="Go" />
- </form>
- </div>
-</div>
-<script>$('#searchbox').show(0);</script>
-
-
-
-
-
-
-
-
- </div>
- </div>
- <div class="clearer"></div>
- </div>
- <div class="footer">
- ©2020, Daniel Frisinghelli.
-
- |
- Powered by <a href="http://sphinx-doc.org/">Sphinx 3.2.1</a>
- & <a href="https://github.com/bitprophet/alabaster">Alabaster 0.7.12</a>
-
- |
- <a href="../_sources/modules/pysegcnn.core.constants.rst.txt"
- rel="nofollow">Page source</a>
- </div>
-
-
-
-
- </body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/modules/pysegcnn.core.dataset.html b/docs/_build/html/modules/pysegcnn.core.dataset.html
deleted file mode 100644
index f80679b0ed647d09a4f947089dc1b62aaebc5c97..0000000000000000000000000000000000000000
--- a/docs/_build/html/modules/pysegcnn.core.dataset.html
+++ /dev/null
@@ -1,161 +0,0 @@
-
-<!DOCTYPE html>
-
-<html>
- <head>
- <meta charset="utf-8" />
- <meta name="viewport" content="width=device-width, initial-scale=1.0" />
- <title>pysegcnn.core.dataset — PySegCNN documentation</title>
- <link rel="stylesheet" href="../_static/alabaster.css" type="text/css" />
- <link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
- <script id="documentation_options" data-url_root="../" src="../_static/documentation_options.js"></script>
- <script src="../_static/jquery.js"></script>
- <script src="../_static/underscore.js"></script>
- <script src="../_static/doctools.js"></script>
- <script src="../_static/language_data.js"></script>
- <link rel="index" title="Index" href="../genindex.html" />
- <link rel="search" title="Search" href="../search.html" />
- <link rel="prev" title="pysegcnn.core.constants" href="pysegcnn.core.constants.html" />
-
- <link rel="stylesheet" href="../_static/custom.css" type="text/css" />
-
-
- <meta name="viewport" content="width=device-width, initial-scale=0.9, maximum-scale=0.9" />
-
- </head><body>
-
-
- <div class="document">
- <div class="documentwrapper">
- <div class="bodywrapper">
-
-
- <div class="body" role="main">
-
- <div class="section" id="module-pysegcnn.core.dataset">
-<span id="pysegcnn-core-dataset"></span><h1>pysegcnn.core.dataset<a class="headerlink" href="#module-pysegcnn.core.dataset" title="Permalink to this headline">¶</a></h1>
-<p>Custom dataset classes compliant to the PyTorch standard.</p>
-<p>Each custom dataset should inherit from torch.utils.data.Dataset to benefit
-from the torch.utils.data.DataLoader class, which implements helpful utilities
-during model training.</p>
-<p>For any kind of image-like dataset, inherit the ImageDataset class to create
-your custom dataset.</p>
-<div class="section" id="license">
-<h2>License<a class="headerlink" href="#license" title="Permalink to this headline">¶</a></h2>
-<blockquote>
-<div><p>Copyright (c) 2020 Daniel Frisinghelli</p>
-<p>This source code is licensed under the GNU General Public License v3.</p>
-<p>See the LICENSE file in the repository’s root directory.</p>
-</div></blockquote>
-<p class="rubric">Classes</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">Cloud95Dataset</span></code>(root_dir[, use_bands, …])</p></td>
-<td><p>Class for the <a href="#id3"><span class="problematic" id="id4">`Cloud-95`_</span></a> dataset by <a href="#id1"><span class="problematic" id="id2">`</span></a>Mohajerani et al.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">ImageDataset</span></code>(root_dir[, use_bands, …])</p></td>
-<td><p>Base class for multispectral image data.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">ProSnowDataset</span></code>(root_dir[, use_bands, …])</p></td>
-<td><p>Class for the ProSnow datasets.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">ProSnowGarmisch</span></code>(root_dir[, use_bands, …])</p></td>
-<td><p>Class for the ProSnow Garmisch dataset.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">ProSnowObergurgl</span></code>(root_dir[, use_bands, …])</p></td>
-<td><p>Class for the ProSnow Obergurgl dataset.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">SparcsDataset</span></code>(root_dir[, use_bands, …])</p></td>
-<td><p>Class for the <a href="#id5"><span class="problematic" id="id6">`Sparcs`_</span></a> dataset.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">StandardEoDataset</span></code>(root_dir[, use_bands, …])</p></td>
-<td><p>Base class for standard Earth Observation style datasets.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">SupportedDatasets</span></code>(value)</p></td>
-<td><p>Names and corresponding classes of the implemented datasets.</p></td>
-</tr>
-</tbody>
-</table>
-</div>
-</div>
-
-
- </div>
-
- </div>
- </div>
- <div class="sphinxsidebar" role="navigation" aria-label="main navigation">
- <div class="sphinxsidebarwrapper">
-<h1 class="logo"><a href="../index.html">PySegCNN</a></h1>
-
-
-
-
-
-
-
-
-<h3>Navigation</h3>
-<p class="caption"><span class="caption-text">Table of contents:</span></p>
-<ul class="current">
-<li class="toctree-l1 current"><a class="reference internal" href="../modules.html">Modules</a><ul class="current">
-<li class="toctree-l2"><a class="reference internal" href="pysegcnn.core.cli.html">pysegcnn.core.cli</a></li>
-<li class="toctree-l2"><a class="reference internal" href="pysegcnn.core.constants.html">pysegcnn.core.constants</a></li>
-<li class="toctree-l2 current"><a class="current reference internal" href="#">pysegcnn.core.dataset</a></li>
-</ul>
-</li>
-</ul>
-
-<div class="relations">
-<h3>Related Topics</h3>
-<ul>
- <li><a href="../index.html">Documentation overview</a><ul>
- <li><a href="../modules.html">Modules</a><ul>
- <li>Previous: <a href="pysegcnn.core.constants.html" title="previous chapter">pysegcnn.core.constants</a></li>
- </ul></li>
- </ul></li>
-</ul>
-</div>
-<div id="searchbox" style="display: none" role="search">
- <h3 id="searchlabel">Quick search</h3>
- <div class="searchformwrapper">
- <form class="search" action="../search.html" method="get">
- <input type="text" name="q" aria-labelledby="searchlabel" />
- <input type="submit" value="Go" />
- </form>
- </div>
-</div>
-<script>$('#searchbox').show(0);</script>
-
-
-
-
-
-
-
-
- </div>
- </div>
- <div class="clearer"></div>
- </div>
- <div class="footer">
- ©2020, Daniel Frisinghelli.
-
- |
- Powered by <a href="http://sphinx-doc.org/">Sphinx 3.2.1</a>
- & <a href="https://github.com/bitprophet/alabaster">Alabaster 0.7.12</a>
-
- |
- <a href="../_sources/modules/pysegcnn.core.dataset.rst.txt"
- rel="nofollow">Page source</a>
- </div>
-
-
-
-
- </body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/objects.inv b/docs/_build/html/objects.inv
index 3e4fffbf2f1c6de6f65ee83cfcdc7726224a294e..79d2d2a07a6c0f2a6e41a623d9c9a89486f6cb12 100644
Binary files a/docs/_build/html/objects.inv and b/docs/_build/html/objects.inv differ
diff --git a/docs/_build/html/py-modindex.html b/docs/_build/html/py-modindex.html
deleted file mode 100644
index 856c1bb9aff4d406a37adf59cc90b2e41352dabd..0000000000000000000000000000000000000000
--- a/docs/_build/html/py-modindex.html
+++ /dev/null
@@ -1,227 +0,0 @@
-
-
-<!DOCTYPE html>
-<html class="writer-html5" lang="en" >
-<head>
- <meta charset="utf-8">
-
- <meta name="viewport" content="width=device-width, initial-scale=1.0">
-
- <title>Python Module Index — PySegCNN documentation</title>
-
-
-
- <link rel="stylesheet" href="_static/css/theme.css" type="text/css" />
- <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
-
-
-
-
-
-
-
- <!--[if lt IE 9]>
- <script src="_static/js/html5shiv.min.js"></script>
- <![endif]-->
-
-
- <script type="text/javascript" id="documentation_options" data-url_root="./" src="_static/documentation_options.js"></script>
- <script src="_static/jquery.js"></script>
- <script src="_static/underscore.js"></script>
- <script src="_static/doctools.js"></script>
- <script src="_static/language_data.js"></script>
-
- <script type="text/javascript" src="_static/js/theme.js"></script>
-
-
- <link rel="index" title="Index" href="genindex.html" />
- <link rel="search" title="Search" href="search.html" />
-
-
-
-</head>
-
-<body class="wy-body-for-nav">
-
-
- <div class="wy-grid-for-nav">
-
- <nav data-toggle="wy-nav-shift" class="wy-nav-side">
- <div class="wy-side-scroll">
- <div class="wy-side-nav-search" >
-
-
-
- <a href="index.html" class="icon icon-home" alt="Documentation Home"> PySegCNN
-
-
-
- </a>
-
-
-
-
-
-
-
-<div role="search">
- <form id="rtd-search-form" class="wy-form" action="search.html" method="get">
- <input type="text" name="q" placeholder="Search docs" />
- <input type="hidden" name="check_keywords" value="yes" />
- <input type="hidden" name="area" value="default" />
- </form>
-</div>
-
-
- </div>
-
-
- <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
-
-
-
-
-
-
- <ul>
-<li class="toctree-l1"><a class="reference internal" href="source/requirements.html">Requirements</a></li>
-<li class="toctree-l1"><a class="reference internal" href="source/installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="source/api.html">API Reference</a></li>
-</ul>
-
-
-
- </div>
-
- </div>
- </nav>
-
- <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
-
-
- <nav class="wy-nav-top" aria-label="top navigation">
-
- <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
- <a href="index.html">PySegCNN</a>
-
- </nav>
-
-
- <div class="wy-nav-content">
-
- <div class="rst-content">
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-<div role="navigation" aria-label="breadcrumbs navigation">
-
- <ul class="wy-breadcrumbs">
-
- <li><a href="index.html" class="icon icon-home"></a> »</li>
-
- <li>Python Module Index</li>
-
-
- <li class="wy-breadcrumbs-aside">
-
- </li>
-
- </ul>
-
-
- <hr/>
-</div>
- <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
- <div itemprop="articleBody">
-
-
- <h1>Python Module Index</h1>
-
- <div class="modindex-jumpbox">
- <a href="#cap-p"><strong>p</strong></a>
- </div>
-
- <table class="indextable modindextable">
- <tr class="pcap"><td></td><td> </td><td></td></tr>
- <tr class="cap" id="cap-p"><td></td><td>
- <strong>p</strong></td><td></td></tr>
- <tr>
- <td><img src="_static/minus.png" class="toggler"
- id="toggle-1" style="display: none" alt="-" /></td>
- <td>
- <code class="xref">pysegcnn</code></td><td>
- <em></em></td></tr>
- <tr class="cg-1">
- <td></td>
- <td>   
- <a href="source/generated/pysegcnn.core.layers.html#module-pysegcnn.core.layers"><code class="xref">pysegcnn.core.layers</code></a></td><td>
- <em></em></td></tr>
- <tr class="cg-1">
- <td></td>
- <td>   
- <a href="source/generated/pysegcnn.core.models.html#module-pysegcnn.core.models"><code class="xref">pysegcnn.core.models</code></a></td><td>
- <em></em></td></tr>
- </table>
-
-
- </div>
-
- </div>
- <footer>
-
-
- <hr/>
-
- <div role="contentinfo">
- <p>
-
- © Copyright 2020, Daniel Frisinghelli
-
- </p>
- </div>
-
-
-
- Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a
-
- <a href="https://github.com/rtfd/sphinx_rtd_theme">theme</a>
-
- provided by <a href="https://readthedocs.org">Read the Docs</a>.
-
-</footer>
-
- </div>
- </div>
-
- </section>
-
- </div>
-
-
- <script type="text/javascript">
- jQuery(function () {
- SphinxRtdTheme.Navigation.enable(true);
- });
- </script>
-
-
-
-
-
-
-</body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/readme.html b/docs/_build/html/readme.html
deleted file mode 100644
index 6e6adb7bff4e577fc262f79a0549041d5ea71569..0000000000000000000000000000000000000000
--- a/docs/_build/html/readme.html
+++ /dev/null
@@ -1,209 +0,0 @@
-
-
-<!DOCTYPE html>
-<html class="writer-html5" lang="en" >
-<head>
- <meta charset="utf-8">
-
- <meta name="viewport" content="width=device-width, initial-scale=1.0">
-
- <title>Test — PySegCNN documentation</title>
-
-
-
- <link rel="stylesheet" href="_static/css/theme.css" type="text/css" />
- <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
-
-
-
-
-
-
-
- <!--[if lt IE 9]>
- <script src="_static/js/html5shiv.min.js"></script>
- <![endif]-->
-
-
- <script type="text/javascript" id="documentation_options" data-url_root="./" src="_static/documentation_options.js"></script>
- <script src="_static/jquery.js"></script>
- <script src="_static/underscore.js"></script>
- <script src="_static/doctools.js"></script>
- <script src="_static/language_data.js"></script>
-
- <script type="text/javascript" src="_static/js/theme.js"></script>
-
-
- <link rel="index" title="Index" href="genindex.html" />
- <link rel="search" title="Search" href="search.html" />
- <link rel="prev" title="<no title>" href="index.html" />
-</head>
-
-<body class="wy-body-for-nav">
-
-
- <div class="wy-grid-for-nav">
-
- <nav data-toggle="wy-nav-shift" class="wy-nav-side">
- <div class="wy-side-scroll">
- <div class="wy-side-nav-search" >
-
-
-
- <a href="index.html" class="icon icon-home" alt="Documentation Home"> PySegCNN
-
-
-
- </a>
-
-
-
-
-
-
-
-<div role="search">
- <form id="rtd-search-form" class="wy-form" action="search.html" method="get">
- <input type="text" name="q" placeholder="Search docs" />
- <input type="hidden" name="check_keywords" value="yes" />
- <input type="hidden" name="area" value="default" />
- </form>
-</div>
-
-
- </div>
-
-
- <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
-
-
-
-
-
-
- <ul class="current">
-<li class="toctree-l1 current"><a class="current reference internal" href="#">Test</a></li>
-</ul>
-
-
-
- </div>
-
- </div>
- </nav>
-
- <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
-
-
- <nav class="wy-nav-top" aria-label="top navigation">
-
- <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
- <a href="index.html">PySegCNN</a>
-
- </nav>
-
-
- <div class="wy-nav-content">
-
- <div class="rst-content">
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-<div role="navigation" aria-label="breadcrumbs navigation">
-
- <ul class="wy-breadcrumbs">
-
- <li><a href="index.html" class="icon icon-home"></a> »</li>
-
- <li>Test</li>
-
-
- <li class="wy-breadcrumbs-aside">
-
-
- <a href="_sources/readme.rst.txt" rel="nofollow"> View page source</a>
-
-
- </li>
-
- </ul>
-
-
- <hr/>
-</div>
- <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
- <div itemprop="articleBody">
-
- <div class="section" id="test">
-<h1>Test<a class="headerlink" href="#test" title="Permalink to this headline">¶</a></h1>
-</div>
-
-
- </div>
-
- </div>
- <footer>
-
- <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
-
-
- <a href="index.html" class="btn btn-neutral float-left" title="<no title>" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left"></span> Previous</a>
-
- </div>
-
-
- <hr/>
-
- <div role="contentinfo">
- <p>
-
- © Copyright 2020, Daniel Frisinghelli
-
- </p>
- </div>
-
-
-
- Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a
-
- <a href="https://github.com/rtfd/sphinx_rtd_theme">theme</a>
-
- provided by <a href="https://readthedocs.org">Read the Docs</a>.
-
-</footer>
-
- </div>
- </div>
-
- </section>
-
- </div>
-
-
- <script type="text/javascript">
- jQuery(function () {
- SphinxRtdTheme.Navigation.enable(true);
- });
- </script>
-
-
-
-
-
-
-</body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/searchindex.js b/docs/_build/html/searchindex.js
index 9287d508bccdc684e700fd3060e6c75678d7f34d..121f50fd6d65dbf40b337319ced81e981d6a4381 100644
--- a/docs/_build/html/searchindex.js
+++ b/docs/_build/html/searchindex.js
@@ -1 +1 @@
-Search.setIndex({docnames:["index","source/api","source/generated/pysegcnn.core.dataset.Cloud95Dataset","source/generated/pysegcnn.core.dataset.ImageDataset","source/generated/pysegcnn.core.dataset.SparcsDataset","source/generated/pysegcnn.core.layers.Block","source/generated/pysegcnn.core.layers.Conv2dSame","source/generated/pysegcnn.core.layers.ConvBnReluMaxPool","source/generated/pysegcnn.core.layers.ConvBnReluMaxUnpool","source/generated/pysegcnn.core.layers.ConvBnReluUpsample","source/generated/pysegcnn.core.layers.Decoder","source/generated/pysegcnn.core.layers.DecoderBlock","source/generated/pysegcnn.core.layers.Encoder","source/generated/pysegcnn.core.layers.EncoderBlock","source/generated/pysegcnn.core.models.Network","source/generated/pysegcnn.core.models.UNet","source/installation","source/requirements"],envversion:{"sphinx.domains.c":2,"sphinx.domains.changeset":1,"sphinx.domains.citation":1,"sphinx.domains.cpp":3,"sphinx.domains.index":1,"sphinx.domains.javascript":2,"sphinx.domains.math":2,"sphinx.domains.python":2,"sphinx.domains.rst":2,"sphinx.domains.std":1,"sphinx.ext.viewcode":1,sphinx:56},filenames:["index.rst","source\\api.rst","source\\generated\\pysegcnn.core.dataset.Cloud95Dataset.rst","source\\generated\\pysegcnn.core.dataset.ImageDataset.rst","source\\generated\\pysegcnn.core.dataset.SparcsDataset.rst","source\\generated\\pysegcnn.core.layers.Block.rst","source\\generated\\pysegcnn.core.layers.Conv2dSame.rst","source\\generated\\pysegcnn.core.layers.ConvBnReluMaxPool.rst","source\\generated\\pysegcnn.core.layers.ConvBnReluMaxUnpool.rst","source\\generated\\pysegcnn.core.layers.ConvBnReluUpsample.rst","source\\generated\\pysegcnn.core.layers.Decoder.rst","source\\generated\\pysegcnn.core.layers.DecoderBlock.rst","source\\generated\\pysegcnn.core.layers.Encoder.rst","source\\generated\\pysegcnn.core.layers.EncoderBlock.rst","source\\generated\\pysegcnn.core.models.Network.rst","source\\generated\\pysegcnn.core.models.UNet.rst","source\\installation.rst","source\\requirements.rst"],objects:{"pysegcnn.core.dataset":{Cloud95Dataset:[2,0,1,""],ImageDataset:[3,0,1,""],SparcsDataset:[4,0,1,""]},"pysegcnn.core.dataset.Cloud95Dataset":{__init__:[2,1,1,""]},"pysegcnn.core.dataset.ImageDataset":{__init__:[3,1,1,""]},"pysegcnn.core.dataset.SparcsDataset":{__init__:[4,1,1,""]},"pysegcnn.core.layers":{Block:[5,0,1,""],Conv2dSame:[6,0,1,""],ConvBnReluMaxPool:[7,0,1,""],ConvBnReluMaxUnpool:[8,0,1,""],ConvBnReluUpsample:[9,0,1,""],Decoder:[10,0,1,""],DecoderBlock:[11,0,1,""],Encoder:[12,0,1,""],EncoderBlock:[13,0,1,""]},"pysegcnn.core.layers.Block":{__init__:[5,1,1,""]},"pysegcnn.core.layers.Conv2dSame":{__init__:[6,1,1,""]},"pysegcnn.core.layers.ConvBnReluMaxPool":{__init__:[7,1,1,""]},"pysegcnn.core.layers.ConvBnReluMaxUnpool":{__init__:[8,1,1,""]},"pysegcnn.core.layers.ConvBnReluUpsample":{__init__:[9,1,1,""]},"pysegcnn.core.layers.Decoder":{__init__:[10,1,1,""]},"pysegcnn.core.layers.DecoderBlock":{__init__:[11,1,1,""]},"pysegcnn.core.layers.Encoder":{__init__:[12,1,1,""]},"pysegcnn.core.layers.EncoderBlock":{__init__:[13,1,1,""]},"pysegcnn.core.models":{Network:[14,0,1,""],UNet:[15,0,1,""]},"pysegcnn.core.models.Network":{__init__:[14,1,1,""]},"pysegcnn.core.models.UNet":{__init__:[15,1,1,""]}},objnames:{"0":["py","class","Python class"],"1":["py","method","Python method"]},objtypes:{"0":"py:class","1":"py:method"},terms:{"001":[],"08768":[],"10077":[],"1901":[],"2001":[],"2014":[],"2019":[],"2020":2,"2072":[],"2x2":[7,8,15],"4292":[],"4907":[],"case":16,"class":[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15],"default":[2,3,4,16],"enum":[],"final":16,"float":[],"function":1,"import":16,"int":[2,3,4,5,6,7,8,9,10,11,12,13,15],"new":14,"public":0,"return":[2,3,4,5,6,7,8,9,10,11,12,13,14,15],"static":[],"true":[2,3,4,10],"var":[],"while":[0,16],And:17,For:[0,2,3,4],NOT:[],Not:[],The:[0,1,2,3,4,10,12,14,16],These:[],Use:[],Used:[2,3,4],Useful:[2,3,4],__call__:[5,6,7,8,9,10,11,12,13,14,15],__init__:[2,3,4,5,6,7,8,9,10,11,12,13,14,15],_earli:[],_equat:[],_graphic:[],_model:[],_mtl:[],_sampl:[],_subplot:[],_torch:6,abl:16,about:[],abov:16,abs:[],access:[],accord:[],accur:[2,3,4],accuraci:[],accuracy_funct:[],acknowledg:[],activ:16,adam:[],add:16,added:[],addit:[5,6,7,8,9,10,11,12,13,15],adjust:[],advanc:[],aerosol:[],after:16,algorithm:[],all:[2,3,4,16,17],along:[],alpha:[],also:[],altern:16,among:[],amount:[],angl:[],ani:16,api:0,append:[],appli:[10,15],arbitrari:[],architechur:[10,12],architectur:1,archiv:[],arg:6,argument:[5,6,7,8,9,10,11,12,13,15],arrai:[],array_lik:[],artifici:[],arxiv:[],aspx:[],assertionerror:[],assum:[],atmopsher:[],atmospher:[],attribut:[5,6,7,8,9,10,11,12,13,14,15],augment:[2,3,4],autom:[],avail:1,axes:[],axessubplot:[],axi:[],band:[2,3,4],band_1:[],band_2:[],band_n:[],band_name_1:[],band_name_2:[],band_name_n:[],base:[0,3,14],baseconfig:[],bash:[],basic:5,batch:15,batch_siz:[],batchnorm:[7,8,9],befor:[],beginn:[],belong:[],below:16,benefit:[],best:[],better:[],bia:[],blob:[],block:[7,8,9,10,11,12,13,15],blue:[],bolzano:0,bool:[2,3,4,10,15],both:[5,6,7,8,9,10,11,12,13,14,15,17],bottom:[],box:1,bright:[],build:[0,1,16,17],build_sampl:[],cach:[],calcul:6,calibr:[],call:[],can:[16,17],care:17,ccisnow:16,center:[2,3,4],centimet:[],chang:[],channel:[5,6,7,8,9,10,11,12,13,15,16],check:[],check_tile_extend:[],checkpoint:[],checkpoint_st:[],chronolog:[2,3,4],cirru:[],classif:[],classifi:[],clear:[],cli:[],clone:16,cloud95:[],cloud95dataset:[],cloud95label:[],cloud:2,cmap:[],code:[0,16],collect:[],color:[],colormap:[],column:[],com:0,command:16,common:[],complet:[],compliant:[1,3],compos:[2,3,4],compose_scen:[],composit:[],compsit:[],comput:[],concaten:[],conda:17,conf_mat:[],config:16,configur:[],confus:[],connect:[10,15],consid:[],constant:[2,3,4],constrast:[],construct:[],contact:[],contain:[],content:[],contrast:[],contrast_stretch:[],contributor:[],control:[],conv2d:6,conv2dsam:[5,7,8,9,10,11,12,13,15],conv:[],conv_bn_relu:[],convbnrelumaxpool:12,convbnrelumaxunpool:10,convbnreluupsampl:[],convent:[2,3,4],convers:[],convert:[],convolut:[1,5,6,7,8,9,10,11,12,13,15],convolv:6,copi:16,copyright:[],core:[],corner:[],correctli:[],correspond:[2,3,4],counter:[],cpu:[],creat:16,criterion:[],crossentropi:[],crossentropyloss:[],cuda:[],current:16,custom:1,customsubset:[],cval:[],dai:[],daniel:0,darkblu:[],data:[2,3,4],data_loading_tutori:[],dataclass:[],dataload:3,dataset:[],dataset_config:[],dataset_nam:[],datasetconfig:[],date:[],date_scene_split:[],dateformat:[],datesplit:[],datetim:[],decod:[11,12,15],decoderblock:10,decreas:[],deep:0,defaul:[],defin:[10,12,16],degre:[],delta:[],depend:16,depth:[10,12],descent:[],describ:[],descript:[],design:0,destack:[],destack_tiff:[],detail:[],determin:[10,12],develop:0,dict:[5,6,7,8,9,10,11,12,13,15],dictconfig:[],dictionari:[],did:[],digit:[],digiti:[],dilat:6,dimens:6,directli:[],directori:[2,3,4,16],disabl:[],disjoint:[],disk:[],distinct:[2,3,4],distribut:[],divid:[2,3,4],divis:[2,3,4],doc:6,document:17,doe:5,doi:[],download:[],downsampl:[],doy2dat:[],drop:[],dtype:[],dure:3,each:[2,3,4,10,12,14,15],earli:[],early_stop:[],earlystop:[],earth:0,easili:[],edu:[],either:[],element:[10,12],els:[],empti:[],enabl:3,enc_cach:[],encod:[10,13,15],encoderblock:12,end:[],entir:[],enumer:[],env:16,environ:17,epoch:[],equal:[],equat:[],error:[],etc:[],eurac:0,eval:[],eval_config:[],evalconfig:[],evalu:[],evenli:[2,3,4],exampl:[10,12],exce:[],except:16,exclud:[],exist:[],explicitli:0,express:[2,3,4],extend:[],extens:[],extra:[],extra_repr:[],extract:[],extract_arch:[],factor:[],factori:[],fals:[2,3,4],fcc:[],featur:[],feed:[],fifth:[],fig:[],figsiz:[],figur:[],file:[2,3,4,14,15,16,17],filenam:[],filenotfounderror:[],filter:[10,12,15],first:[10,12,16],fit:[],flip:[],fliplr:[],flipud:[],flood:[],folder:[],follow:[1,15,16],forg:16,format:[],forward:6,found:[],four:[],fraction:[],framework:0,freez:[],frisinghelli:0,from:[3,10,15,16],fulli:0,further:0,furthermor:3,futur:0,garmisch:[],gaussian:[],gdal:17,gener:[0,1,2,3,4,6,10,12,14],get:[],get_label:[],get_num_thread:[],get_radiometric_const:[],get_sensor:[],get_siz:[],git:16,github:[],gitlab:16,given:6,gmail:0,gnu:0,gov:[],gradient:[],graphic:[],greater:[],green:[],grei:[],ground:[2,3,4],group:[],gt_pattern:[2,3,4],guid:16,handi:3,handl:[],hardwar:16,has:[6,10,12],have:16,hay:[],heavili:16,height:[],help:[2,3,4],here:[16,17],horizont:[],host:[],how:[],howev:[],html:6,http:[6,16],hugh:[],idea:0,identifi:[],ids:[],idx:[],imag:[2,3,4,10,12,15],imagedataset:[],imageri:0,img2np:[],img_siz:[],implement:[1,10,12,14,15],improv:[],in_channel:[5,6,7,8,9,11,13,15],includ:[],increas:[],index:[],indic:[],inf:16,infer:[],inform:0,infrar:[],inherit:[3,14],init_dataset:[],init_log:[],init_loss_funct:[],init_model:[],init_optim:[],init_st:[],init_str:[],initi:[2,3,4,5,6,7,8,9,10,11,12,13,14,15],initili:[],inpath:[],input:[2,3,4,5,6,7,8,9,10,11,12,13,15],instal:[0,17],instanc:[2,3,4,5],instanci:[10,12],instead:[],institut:0,instruct:16,instrument:[],integ:[],interfac:[],intermedi:[],intern:[5,6,7,8,9,10,11,12,13,14,15],interpol:[],interpret:16,invari:[],invarianttransform:[],is_divis:[],item:[2,3,4],item_in_enum:[],its:6,kei:[],kernel:6,kernel_s:6,keyword:[6,7,8,15],kind:[],kwarg:[5,6,7,8,9,10,11,12,13,15],label:[2,3,4],land:[],landsat8:[],landsat:[],landsat_radiometric_calibr:[],last:[],later:[],latest:[],launch:[],layer:15,learn:0,left:[],len:[],length:[10,12],level:[],licens:[],lightblu:[],lightgreen:[],like:[],line:16,linear:[],linux:17,list:[1,2,3,4,6,10,12,15,17],load:[],load_checkpoint:[],loader:[],local:[],log:[],log_conf:[],logconfig:[],logfil:[],logging_config:[],loss:[],loss_funct:[],loss_nam:[],machin:[0,16],mai:16,main:[],mainli:[],make:[],manag:[],master:[],match:[2,3,4],matplotlib:17,matrix:[],max:[7,8,15],max_pool2d:[],max_unpool2d:[],maxima:[],maximum:[],mdpi:[],mean:[],measur:[],met:[],metadata:[],method:[2,3,4,5,6,7,8,9,10,11,12,13,14,15],metric:[],min:[],min_delta:[],mini:[],miniconda:[],minimum:[],mode:[],model:[3,10,12],model_config:[],model_dataset_optimizer_splitmode_splitparams_tilesize_batchsize_band:[],model_nam:[],model_st:[],modelconfig:[],modifi:15,modul:[5,6,7,8,9,10,11,12,13,14,15],mohajerani:2,more:[],mountain:[],move:[],msi:[],multispectr:[0,1,3],name:[2,3,4,16],navig:16,nclass:15,ncol:[],ndarrai:[],nearest:9,need:[],neighbor:9,nest:[],net:15,network:[],networktrain:[],neural:[],nir:[],nli:[],nnir:[],nois:[],none:[2,3,4,5,6,7,8,9,10,11,12,13,14,15],norm:[],normal:15,note:[],notimplementederror:[],now:[],nrow:[],nthread:[],ntile:[],number:[5,6,7,8,9,10,11,12,13,15],numpi:17,obergurgl:[],object:[],observ:0,offici:16,oli:[],onc:16,one:[],onli:[2,3,4],open:1,oper:15,optim:[],optim_nam:[],option:[2,3,4],order:[],org:6,origin:[2,3,4],otherwis:0,out:1,out_channel:[5,6,7,8,9,11,13],outpath:[],output:[5,6,7,8,9,11,13],output_pad:[],over:[],overwrit:[],overwritten:[],packag:[0,17],pad:[2,3,4,6],padding_mod:[],page:1,pair:[],pairwis:[],pairwise_disjoint:[],pan:[],paramet:[2,3,4,5,6,7,8,9,10,11,12,13,15],pars:[],parse_landsat_scen:[],parse_scene_id:[],parse_sentinel2_scen:[],parser:[],part:0,pass:[5,6,7,8,9,10,11,12,13,15],path:[2,3,4],pathlib:[],patienc:[],pattern:[],percentil:[],perform:[],pip:16,pixel:[],plane:[],pleas:0,plot:[],plot_band:[],plot_confusion_matrix:[],plot_loss:[],plot_sampl:[],plot_scen:[],point:[],pool:7,posit:6,possibl:[],precis:[],predict:[],predict_sampl:[],predict_scen:[],preprocess:[],preserv:6,pretrain:[],pretrained_model:[],previou:[],primarili:0,print:[],probabl:[],procedur:[],process:[],product:[],project:[],propag:[],properti:[],prosnow:[],prosnowdataset:[],prosnowgarmisch:[],prosnowlabel:[],prosnowobergurgl:[],provid:16,publicli:[],pure:17,pyplot:[],pysegcnn:[1,17],pytest:17,python:[16,17],pytorch:[0,1,3,6,15,17],quit:[],radianc:[],radiometr:[],rais:5,random:[2,3,4],random_scene_split:[],random_tile_split:[],randomli:[],randomscenesplit:[],randomsplit:[],randomsubset:[],randomtilesplit:[],rate:[],ratio:[],raw:[],read:[],read_landsat_metadata:[],read_scen:[],readm:[],recal:[],recommend:16,reconstruct:[],reconstruct_scen:[],rectifi:[],red:[],refer:0,reflect:[],regur:[2,3,4],relu:[7,8,9],remain:[],remot:[],remov:[],remove_raw:[],remsen:16,repositori:[0,16],repres:[],represent:[],reproduc:[2,3,4],requir:[0,16],rescal:[],research:0,reset_paramet:6,reshap:[],resourc:[],respect:[],restructur:[],result:[],resum:[],retriev:[],right:[],root:[2,3,4,16],root_dir:[2,3,4],rotat:[],row:[],run:16,running_mean:[],saeedi:2,same:[6,10,12],same_pad:[],sampl:[2,3,4,16],satellit:0,save:[],save_st:[],scene:[2,3,4],scene_data:[],scene_id:[],scene_id_1:[],scene_id_1_b1:[],scene_id_1_b2:[],scene_id_1_bn:[],scene_id_2:[],scene_id_2_b1:[],scene_id_2_b2:[],scene_id_2_bn:[],scene_id_n:[],scenesubset:[],score:[],scratch:[],script:[],scriptmodul:[5,6,7,8,9,10,11,12,13,14,15],search:[2,3,4],see:[0,2,3,4,10,12,17],seed:[2,3,4],segment:[],self:[2,3,4],semant:0,sens:[],sensor:[],sentinel2:[],sentinel:[],sequenc:[],sequenti:5,seri:[2,3,4],set:[2,3,4],shadow:[],shadow_over_wat:[],shape:[2,3,4,6],share:[5,6,7,8,9,10,11,12,13,14,15],share_memori:[5,6,7,8,9,10,11,12,13,14,15],ship:17,should:[10,12,16],show:[],sienna:[],signatur:[2,3,4],simpli:[],singl:[],size:[2,3,4,6],skip:[10,15],skip_connect:[],skyblu:[],slightli:15,snow:[],snow_fre:[],some:16,sorourmo:[],sort:[2,3,4],sourc:[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16],source_path:[],spaceborn:1,sparc:4,sparcsdataset:[],sparcslabel:[],spatial:6,specif:16,specifi:[],speckl:[],spectral:[2,3,4],sphinx:17,sphinx_rtd_them:17,split:[2,3,4],split_config:[],split_mod:[],splitconfig:[],squar:[2,3,4],stabl:6,stack:[],standard:[1,3],standard_eo_structur:[],standardeodataset:[],start:[],state:[0,5,6,7,8,9,10,11,12,13,14,15],state_fil:[],stateconfig:[],steelblu:[],step:[],stop:[],store:[],str:[2,3,4,5,6,7,8,9,10,11,12,13,15],stretch:[],stride:6,string:[],strptime:[],structur:[],structure_pars:[],style:[],sub:[],subclass:[10,12],subdir:[],subdirectori:[2,3,4],submodul:[],subset:[],subset_typ:[],success:16,successfulli:16,suffix:[],sum:[],support:16,supporteddataset:[],supportedlossfunct:[],supportedmodel:[],supportedoptim:[],supportedsplit:[],sure:[],swir1:[],swir2:[],system:16,tabl:[],take:[16,17],target_path:[],task:[],temperatur:[],tensor:[],termin:16,test:[2,3,4],test_d:[],test_dl:[],text:[],than:[],thei:[],therefor:16,thermal:[],thi:[0,1,16],thread:[],threshold:[],tif:[2,3,4],tiff:[],tile:[2,3,4],tile_00:[],tile_01:[],tile_0n:[],tile_10:[],tile_11:[],tile_1n:[],tile_m0:[],tile_m1:[],tile_mn:[],tile_s:[2,3,4],tile_topleft_corn:[],time:[2,3,4],tir1:[],tir:[],to_tensor:[],toa_brt:[],toa_rad:[],toa_ref:[],togeth:[2,3,4],top:[],topleft:[],torch:[3,5,6],torch_se:[],total:[2,3,4],train:[2,3,4],train_d:[],train_dl:[],train_val_test_split:[],trainer:[],training_st:[],transfer:[],transfer_model:[],transform:[2,3,4],transpos:[],tree:[],truth:[2,3,4],ttratio:[],tupl:6,tutori:[],tvratio:[],two:[],txt:[],type:[2,3,4,5,6,7,8,9,10,11,12,13,14,15,16],typeerror:5,under:0,unet:[10,12],unfreez:[],unibz:16,union:[],unit:[],unpool:[8,15],untransform:[2,3,4],updat:[],upsampl:[9,15],usag:[],use:[2,3,4,16,17],use_band:[2,3,4],used:[2,3,4],useful:[],user:[],usg:[],using:[3,16,17],util:3,valid:[2,3,4],valid_d:[],valid_dl:[],valu:[2,3,4],valueerror:[],vapor:[],varianc:[],variant:[],varianttransform:[],verif:[],verifi:16,version:[2,3,4,15,16],vertic:[],violet:[],vnir1:[],vnir2:[],vnir3:[],wait:[],walkthrough:[],want:[14,17],water:[],websit:16,weight:[],when:[10,12],where:[],whether:[2,3,4,10,15],which:[],white:[],width:[],wiki:[],wikipedia:[],window:17,within:16,work:[0,17],worst:[],wrap:[],wrapper:[],write:[],wrote:0,www:[],y_pred:[],year:[],yellow:[],yml:[16,17],you:[14,16,17],your:16,yyyymmdd:[],zero:[2,3,4]},titles:["PySegCNN: Image segmentation with convolutional neural networks in Python","API Reference","pysegcnn.core.dataset.Cloud95Dataset","pysegcnn.core.dataset.ImageDataset","pysegcnn.core.dataset.SparcsDataset","pysegcnn.core.layers.Block","pysegcnn.core.layers.Conv2dSame","pysegcnn.core.layers.ConvBnReluMaxPool","pysegcnn.core.layers.ConvBnReluMaxUnpool","pysegcnn.core.layers.ConvBnReluUpsample","pysegcnn.core.layers.Decoder","pysegcnn.core.layers.DecoderBlock","pysegcnn.core.layers.Encoder","pysegcnn.core.layers.EncoderBlock","pysegcnn.core.models.Network","pysegcnn.core.models.UNet","Installation","Requirements"],titleterms:{Using:0,acknowledg:0,api:1,architechtur:1,block:5,cli:[],cloud95dataset:2,conda:16,config:[],constant:[],contact:0,content:[],contributor:0,conv2dsam:6,convbnrelumaxpool:7,convbnrelumaxunpool:8,convbnreluupsampl:9,convolut:0,core:[2,3,4,5,6,7,8,9,10,11,12,13,14,15],dataset:[1,2,3,4],decod:[1,10],decoderblock:11,depend:17,document:[],download:16,encod:[1,12],encoderblock:13,environ:16,eval:[],framework:17,graphic:[],imag:[0,1],imagedataset:3,instal:16,layer:[1,5,6,7,8,9,10,11,12,13],learn:17,licens:0,log:[],machin:17,main:[],manag:16,model:[1,14,15],modul:[],network:[0,1,14],neural:[0,1],packag:16,predict:[],preprocess:[],pysegcnn:[0,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16],python:0,pytorch:16,raster:17,readm:[],refer:1,requir:17,scipi:17,segment:0,sparc:[],sparcsdataset:4,split:[],stack:17,submodul:[],support:1,test:17,tool:17,train:[],trainer:[],transform:[],unet:15,util:[]}})
\ No newline at end of file
+Search.setIndex({docnames:["index","source/api","source/generated/pysegcnn.core.dataset.Cloud95Dataset","source/generated/pysegcnn.core.dataset.ImageDataset","source/generated/pysegcnn.core.dataset.SparcsDataset","source/generated/pysegcnn.core.dataset.StandardEoDataset","source/generated/pysegcnn.core.layers.Block","source/generated/pysegcnn.core.layers.Conv2dSame","source/generated/pysegcnn.core.layers.ConvBnReluMaxPool","source/generated/pysegcnn.core.layers.ConvBnReluMaxUnpool","source/generated/pysegcnn.core.layers.ConvBnReluUpsample","source/generated/pysegcnn.core.layers.Decoder","source/generated/pysegcnn.core.layers.DecoderBlock","source/generated/pysegcnn.core.layers.Encoder","source/generated/pysegcnn.core.layers.EncoderBlock","source/generated/pysegcnn.core.models.Network","source/generated/pysegcnn.core.models.UNet","source/generated/pysegcnn.core.transforms.Augment","source/installation","source/requirements"],envversion:{"sphinx.domains.c":2,"sphinx.domains.changeset":1,"sphinx.domains.citation":1,"sphinx.domains.cpp":3,"sphinx.domains.index":1,"sphinx.domains.javascript":2,"sphinx.domains.math":2,"sphinx.domains.python":2,"sphinx.domains.rst":2,"sphinx.domains.std":1,"sphinx.ext.viewcode":1,sphinx:56},filenames:["index.rst","source\\api.rst","source\\generated\\pysegcnn.core.dataset.Cloud95Dataset.rst","source\\generated\\pysegcnn.core.dataset.ImageDataset.rst","source\\generated\\pysegcnn.core.dataset.SparcsDataset.rst","source\\generated\\pysegcnn.core.dataset.StandardEoDataset.rst","source\\generated\\pysegcnn.core.layers.Block.rst","source\\generated\\pysegcnn.core.layers.Conv2dSame.rst","source\\generated\\pysegcnn.core.layers.ConvBnReluMaxPool.rst","source\\generated\\pysegcnn.core.layers.ConvBnReluMaxUnpool.rst","source\\generated\\pysegcnn.core.layers.ConvBnReluUpsample.rst","source\\generated\\pysegcnn.core.layers.Decoder.rst","source\\generated\\pysegcnn.core.layers.DecoderBlock.rst","source\\generated\\pysegcnn.core.layers.Encoder.rst","source\\generated\\pysegcnn.core.layers.EncoderBlock.rst","source\\generated\\pysegcnn.core.models.Network.rst","source\\generated\\pysegcnn.core.models.UNet.rst","source\\generated\\pysegcnn.core.transforms.Augment.rst","source\\installation.rst","source\\requirements.rst"],objects:{"pysegcnn.core.dataset":{Cloud95Dataset:[2,0,1,""],ImageDataset:[3,0,1,""],SparcsDataset:[4,0,1,""],StandardEoDataset:[5,0,1,""]},"pysegcnn.core.dataset.Cloud95Dataset":{__init__:[2,1,1,""]},"pysegcnn.core.dataset.ImageDataset":{__init__:[3,1,1,""]},"pysegcnn.core.dataset.SparcsDataset":{__init__:[4,1,1,""]},"pysegcnn.core.dataset.StandardEoDataset":{__init__:[5,1,1,""]},"pysegcnn.core.layers":{Block:[6,0,1,""],Conv2dSame:[7,0,1,""],ConvBnReluMaxPool:[8,0,1,""],ConvBnReluMaxUnpool:[9,0,1,""],ConvBnReluUpsample:[10,0,1,""],Decoder:[11,0,1,""],DecoderBlock:[12,0,1,""],Encoder:[13,0,1,""],EncoderBlock:[14,0,1,""]},"pysegcnn.core.layers.Block":{__init__:[6,1,1,""]},"pysegcnn.core.layers.Conv2dSame":{__init__:[7,1,1,""]},"pysegcnn.core.layers.ConvBnReluMaxPool":{__init__:[8,1,1,""]},"pysegcnn.core.layers.ConvBnReluMaxUnpool":{__init__:[9,1,1,""]},"pysegcnn.core.layers.ConvBnReluUpsample":{__init__:[10,1,1,""]},"pysegcnn.core.layers.Decoder":{__init__:[11,1,1,""]},"pysegcnn.core.layers.DecoderBlock":{__init__:[12,1,1,""]},"pysegcnn.core.layers.Encoder":{__init__:[13,1,1,""]},"pysegcnn.core.layers.EncoderBlock":{__init__:[14,1,1,""]},"pysegcnn.core.models":{Network:[15,0,1,""],UNet:[16,0,1,""]},"pysegcnn.core.models.Network":{__init__:[15,1,1,""]},"pysegcnn.core.models.UNet":{__init__:[16,1,1,""]},"pysegcnn.core.transforms":{Augment:[17,0,1,""]},"pysegcnn.core.transforms.Augment":{__init__:[17,1,1,""]}},objnames:{"0":["py","class","Python class"],"1":["py","method","Python method"]},objtypes:{"0":"py:class","1":"py:method"},terms:{"1x1":16,"2014":4,"2020":2,"2x2":[8,9,16],"case":18,"class":[2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17],"default":[2,3,4,5,18],"final":18,"function":1,"import":18,"int":[2,3,4,5,6,7,8,9,10,11,12,13,14,16],"new":15,"public":0,"return":[6,8,9,10,12,14,17],"true":[2,3,4,5,11],"while":[0,18],And:19,For:[0,2,3,4,5],The:[0,2,3,4,5,6,7,11,13,15,16,18],Used:[2,3,4,5],Useful:[2,3,4,5],__init__:[2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17],abl:18,abov:[5,18],accur:17,activ:18,add:18,addit:[6,7,8,9,10,11,12,13,14,16],after:18,all:[2,3,4,5,18,19],altern:18,amount:7,ani:18,api:0,appli:[11,16,17],architechur:[11,13],architectur:1,arg:7,argument:[6,7,8,9,10,11,12,13,14,16],attribut:[6,7,8,9,10,11,12,13,14,15,16],augment:[2,3,4,5],avail:1,band:[2,3,4,5],base:[0,3,5,15],basic:6,batch:16,batchnorm:[8,9,10],below:18,block:[8,9,10,11,12,13,14,16],bolzano:0,bool:[2,3,4,5,11,16],both:19,box:1,build:[0,1,18,19],cach:13,calcul:7,can:[5,18,19],care:19,ccisnow:18,center:[2,3,4,5],channel:[6,7,8,9,10,11,12,13,14,16,18],chronolog:[2,3,4,5],classif:16,classifi:16,clone:18,cloud:2,code:[0,18],com:0,command:18,compliant:[1,3],compos:[2,3,4,5],compose_scen:5,conda:19,config:18,connect:[11,16],constant:[2,3,4,5],contain:17,conv2d:7,conv2dsam:[6,8,9,10,11,12,13,14,16],conv:6,convbnrelumaxpool:[11,13],convbnrelumaxunpool:11,convent:[2,3,4,5],convolut:[1,6,7,8,9,10,11,12,13,14,16],convolv:7,copi:18,correspond:[2,3,4,5],creat:18,current:[1,18],custom:1,daniel:0,data:[2,3,4,5],dataload:3,decod:[12,13,16],decoderblock:11,deep:0,defin:[11,13,18],depend:18,depth:[11,13],design:0,determin:[11,13],develop:0,dict:[6,7,8,9,10,11,12,13,14,16],dictionari:13,dilat:7,dimens:7,directli:5,directori:[1,2,3,4,5,18],distinct:[2,3,4,5],divid:[2,3,4,5],divis:[2,3,4,5],document:19,doe:[6,8,9,10,12,14],dure:3,each:[2,3,4,5,11,13,15,16,17],earth:[0,5],element:[11,13],enabl:3,encod:[11,14,16],encoderblock:[11,13],env:18,environ:19,epoch:16,eurac:0,evenli:[2,3,4,5],exampl:[5,11,13],except:18,explicitli:0,express:[2,3,4,5],fals:[2,3,4,5],featur:[11,13],file:[2,3,4,5,15,18,19],filter:[11,13,16],first:[11,13,18],follow:[1,5,16,18],forg:18,framework:0,frisinghelli:0,from:[3,11,16,18],fulli:0,further:0,futur:0,gdal:19,gener:[0,2,3,4,5,11,13,15],git:18,gitlab:18,given:7,gmail:0,gnu:0,ground:[2,3,4,5],gt_pattern:[2,3,4,5],guid:18,handi:3,hardwar:18,has:[7,11,13],have:18,hay:4,heavili:18,help:17,here:[18,19],http:18,hugh:4,idea:0,imag:[1,2,3,4,5,11,13,16],imageri:0,implement:[1,5,11,13,15,16],in_channel:[6,7,8,9,10,12,14,16],indic:13,inf:18,inform:0,inherit:[3,5,15],initi:[2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17],input:[2,3,4,5,6,7,8,9,10,11,12,13,14,16],instal:[0,19],instanc:[2,3,4,5,6,8,9,10,12,14,17],instanci:[11,13],institut:0,instruct:18,intermedi:13,interpret:18,invarianttransform:17,item:[2,3,4,5],its:7,kei:13,kernel:7,kernel_s:7,keyword:[7,16],kwarg:[6,7,8,9,10,11,12,13,14,16],label:[2,3,4,5],layer:16,learn:0,length:[11,13],line:18,linux:19,list:[1,2,3,4,5,7,11,13,16,17,19],machin:[0,18],mai:18,match:[2,3,4,5],matplotlib:19,max:[8,9,13,16],method:[2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17],model:[3,11,13],modifi:16,modulelist:[11,13],mohajerani:2,multispectr:[0,1,3],name:[2,3,4,5,18],navig:18,nclass:16,ndarrai:[11,13],nearest:10,neighbor:10,net:16,nfilter:16,none:[2,3,4,5,15,17],normal:16,number:[6,7,8,9,10,11,12,13,14,16],numpi:[11,13,19],observ:[0,5],offici:18,onc:18,onli:[2,3,4,5],open:1,oper:16,option:[2,3,4,5,11],order:17,origin:[2,3,4,5],otherwis:0,out:1,out_channel:[6,7,8,9,10,12,14],output:[6,7,8,9,10,12,13,14],packag:[0,19],pad:[2,3,4,5,7],pad_height:7,pad_width:7,page:1,paramet:[2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17],part:0,pass:[6,7,8,9,10,11,12,13,14,16],path:[2,3,4,5,15],pathlib:15,pip:18,pleas:0,pool:[8,13],posit:7,preserv:7,primarili:0,provid:18,pure:19,pysegcnn:[1,19],pytest:19,python:[18,19],pytorch:[0,1,3,16,19],rais:[6,8,9,10,12,14],random:[2,3,4,5],recommend:18,refer:0,regur:[2,3,4,5],relu:[8,9,10],remain:5,remsen:18,repositori:[0,18],reproduc:[2,3,4,5],requir:[0,13,18],research:0,root:[2,3,4,5,18],root_dir:[2,3,4,5],run:18,saeedi:2,same:[7,11,13],sampl:[2,3,4,5,18],satellit:0,save:15,scene:[2,3,4,5],scene_id_1:5,scene_id_1_b1:5,scene_id_1_b2:5,scene_id_1_bn:5,scene_id_2:5,scene_id_2_b1:5,scene_id_2_b2:5,scene_id_2_bn:5,scene_id_n:5,search:[2,3,4,5],see:[0,5,11,13,17,19],seed:[2,3,4,5],self:17,semant:0,sequenc:17,sequenti:[6,8,9,10,12,14],seri:[2,3,4,5],set:[2,3,4,5],shape:[2,3,4,5,7],share:5,ship:19,should:[11,13,18],signatur:17,size:[2,3,4,5,7],skip:[11,16],slightli:16,some:[1,18],sort:[2,3,4,5],sourc:[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18],spaceborn:1,sparc:4,sparcsdataset:5,spatial:7,specif:18,spectral:[2,3,4,5],sphinx:19,sphinx_rtd_them:19,split:[2,3,4,5],squar:[2,3,4,5],standard:[1,3,5],standard_eo_structur:5,state:[0,15],state_fil:15,str:[2,3,4,5,6,7,8,9,10,11,12,13,14,15,16],stride:7,structur:[1,5],style:5,subclass:[11,13],subdirectori:[2,3,4,5],success:18,successfulli:18,support:[1,18],system:18,take:[18,19],tensor:13,termin:18,test:[2,3,4,5],therefor:18,thi:[0,1,3,5,18],tif:[2,3,4,5],tile:[2,3,4,5],tile_s:[2,3,4,5],time:[2,3,4,5],togeth:[2,3,4,5],torch:[3,6,7,8,9,10,11,12,13,14],total:[2,3,4,5],train:[2,3,4,5,16],transfer:5,transform:[2,3,4,5],truth:[2,3,4,5],tupl:[7,17],type:[17,18],typeerror:[6,8,9,10,12,14],under:0,unet:[11,13],unibz:18,unpool:[9,16],untransform:[2,3,4,5],upsampl:[10,16],use:[2,3,4,5,18,19],use_band:[2,3,4,5],used:[2,3,4,5],using:[18,19],util:[3,5],valid:[2,3,4,5],valu:[2,3,4,5],varianttransform:17,verifi:18,version:[2,3,4,5,18],want:[15,19],websit:18,when:[11,13],where:15,whether:[2,3,4,5,11,16],window:19,within:18,work:[0,19],wrote:0,yml:[18,19],you:[5,15,18,19],your:[5,18],zero:[2,3,4,5]},titles:["PySegCNN: Image segmentation with convolutional neural networks in Python","API Reference","pysegcnn.core.dataset.Cloud95Dataset","pysegcnn.core.dataset.ImageDataset","pysegcnn.core.dataset.SparcsDataset","pysegcnn.core.dataset.StandardEoDataset","pysegcnn.core.layers.Block","pysegcnn.core.layers.Conv2dSame","pysegcnn.core.layers.ConvBnReluMaxPool","pysegcnn.core.layers.ConvBnReluMaxUnpool","pysegcnn.core.layers.ConvBnReluUpsample","pysegcnn.core.layers.Decoder","pysegcnn.core.layers.DecoderBlock","pysegcnn.core.layers.Encoder","pysegcnn.core.layers.EncoderBlock","pysegcnn.core.models.Network","pysegcnn.core.models.UNet","pysegcnn.core.transforms.Augment","Installation","Requirements"],titleterms:{"class":1,Using:0,acknowledg:0,api:1,architechtur:1,augment:17,block:6,cloud95dataset:2,conda:18,contact:0,contributor:0,conv2dsam:7,convbnrelumaxpool:8,convbnrelumaxunpool:9,convbnreluupsampl:10,convolut:0,core:[2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17],dataset:[1,2,3,4,5],decod:[1,11],decoderblock:12,depend:19,download:18,encod:[1,13],encoderblock:14,environ:18,framework:19,gener:1,imag:0,imagedataset:3,instal:18,layer:[1,6,7,8,9,10,11,12,13,14],learn:19,licens:0,machin:19,manag:18,model:[1,15,16],network:[0,1,15],neural:[0,1],packag:18,pysegcnn:[0,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18],python:0,pytorch:18,raster:19,refer:1,requir:19,scipi:19,segment:0,sparcsdataset:4,specif:1,stack:19,standardeodataset:5,test:19,tool:19,transform:17,unet:16}})
\ No newline at end of file
diff --git a/docs/_build/html/source/api.html b/docs/_build/html/source/api.html
index 8e5f46e9ea6fef9fc109a6c2fd34959e2e453436..67f2d144b6214590aa39d1d875f58112de6ddb81 100644
--- a/docs/_build/html/source/api.html
+++ b/docs/_build/html/source/api.html
@@ -87,11 +87,12 @@
<li class="toctree-l1"><a class="reference internal" href="installation.html">Installation</a></li>
<li class="toctree-l1 current"><a class="current reference internal" href="#">API Reference</a><ul>
<li class="toctree-l2"><a class="reference internal" href="#dataset">Dataset</a><ul>
-<li class="toctree-l3"><a class="reference internal" href="#image-dataset">Image Dataset</a><ul>
+<li class="toctree-l3"><a class="reference internal" href="#generic-classes">Generic classes</a><ul>
<li class="toctree-l4"><a class="reference internal" href="generated/pysegcnn.core.dataset.ImageDataset.html">pysegcnn.core.dataset.ImageDataset</a></li>
+<li class="toctree-l4"><a class="reference internal" href="generated/pysegcnn.core.dataset.StandardEoDataset.html">pysegcnn.core.dataset.StandardEoDataset</a></li>
</ul>
</li>
-<li class="toctree-l3"><a class="reference internal" href="#supported-datasets">Supported datasets</a><ul>
+<li class="toctree-l3"><a class="reference internal" href="#specific-classes">Specific classes</a><ul>
<li class="toctree-l4"><a class="reference internal" href="generated/pysegcnn.core.dataset.SparcsDataset.html">pysegcnn.core.dataset.SparcsDataset</a></li>
<li class="toctree-l4"><a class="reference internal" href="generated/pysegcnn.core.dataset.Cloud95Dataset.html">pysegcnn.core.dataset.Cloud95Dataset</a></li>
</ul>
@@ -194,9 +195,9 @@
<div class="section" id="dataset">
<h2>Dataset<a class="headerlink" href="#dataset" title="Permalink to this headline">¶</a></h2>
<p>Custom dataset classes compliant to the PyTorch <a class="reference external" href="https://pytorch.org/docs/stable/data.html">standard</a>.</p>
-<div class="section" id="image-dataset">
-<h3>Image Dataset<a class="headerlink" href="#image-dataset" title="Permalink to this headline">¶</a></h3>
-<p>Generic class to implement <a class="reference external" href="https://pytorch.org/tutorials/beginner/data_loading_tutorial.html">custom datasets</a>.</p>
+<div class="section" id="generic-classes">
+<h3>Generic classes<a class="headerlink" href="#generic-classes" title="Permalink to this headline">¶</a></h3>
+<p>Generic class to implement a <a class="reference external" href="https://pytorch.org/tutorials/beginner/data_loading_tutorial.html">custom dataset</a>.</p>
<table class="longtable docutils align-default">
<colgroup>
<col style="width: 10%" />
@@ -208,11 +209,24 @@
</tr>
</tbody>
</table>
+<p>Generic class to implement a custom dataset following a standard directory
+structure.</p>
+<table class="longtable docutils align-default">
+<colgroup>
+<col style="width: 10%" />
+<col style="width: 90%" />
+</colgroup>
+<tbody>
+<tr class="row-odd"><td><p><a class="reference internal" href="generated/pysegcnn.core.dataset.StandardEoDataset.html#pysegcnn.core.dataset.StandardEoDataset" title="pysegcnn.core.dataset.StandardEoDataset"><code class="xref py py-obj docutils literal notranslate"><span class="pre">core.dataset.StandardEoDataset</span></code></a></p></td>
+<td><p>Base class for standard Earth Observation style datasets.</p></td>
+</tr>
+</tbody>
+</table>
</div>
-<div class="section" id="supported-datasets">
-<h3>Supported datasets<a class="headerlink" href="#supported-datasets" title="Permalink to this headline">¶</a></h3>
-<p>The following open-source spaceborne multispectral image datasets are supported
-out-of-the-box:</p>
+<div class="section" id="specific-classes">
+<h3>Specific classes<a class="headerlink" href="#specific-classes" title="Permalink to this headline">¶</a></h3>
+<p>Specific classes for some open-source image datasets. Currently, the following
+spaceborne multispectral image datasets are supported out-of-the-box:</p>
<table class="longtable docutils align-default">
<colgroup>
<col style="width: 10%" />
@@ -220,7 +234,7 @@ out-of-the-box:</p>
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="generated/pysegcnn.core.dataset.SparcsDataset.html#pysegcnn.core.dataset.SparcsDataset" title="pysegcnn.core.dataset.SparcsDataset"><code class="xref py py-obj docutils literal notranslate"><span class="pre">core.dataset.SparcsDataset</span></code></a></p></td>
-<td><p>Class for the <a class="reference external" href="https://www.usgs.gov/land-resources/nli/landsat/spatial-procedures-automated-removal-cloud-and-shadow-sparcs-validation">Sparcs</a> dataset.</p></td>
+<td><p>Class for the <a class="reference external" href="https://www.usgs.gov/land-resources/nli/landsat/spatial-procedures-automated-removal-cloud-and-shadow-sparcs-validation">Sparcs</a> dataset by <a class="reference external" href="https://www.mdpi.com/2072-4292/6/6/4907">Hughes & Hayes (2014)</a>.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="generated/pysegcnn.core.dataset.Cloud95Dataset.html#pysegcnn.core.dataset.Cloud95Dataset" title="pysegcnn.core.dataset.Cloud95Dataset"><code class="xref py py-obj docutils literal notranslate"><span class="pre">core.dataset.Cloud95Dataset</span></code></a></p></td>
<td><p>Class for the <a class="reference external" href="https://github.com/SorourMo/95-Cloud-An-Extension-to-38-Cloud-Dataset">Cloud-95</a> dataset by <a class="reference external" href="https://arxiv.org/abs/2001.08768">Mohajerani & Saeedi (2020)</a>.</p></td>
@@ -253,7 +267,7 @@ out-of-the-box:</p>
<td><p>Block of convolution, batchnorm, relu and 2x2 max unpool.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="generated/pysegcnn.core.layers.ConvBnReluUpsample.html#pysegcnn.core.layers.ConvBnReluUpsample" title="pysegcnn.core.layers.ConvBnReluUpsample"><code class="xref py py-obj docutils literal notranslate"><span class="pre">core.layers.ConvBnReluUpsample</span></code></a></p></td>
-<td><p>Block of convolution, batchnorm, relu and nearest neighbor upsampling.</p></td>
+<td><p>Block of convolution, batchnorm, relu and nearest neighbor upsample.</p></td>
</tr>
</tbody>
</table>
@@ -295,7 +309,7 @@ architecture.</p>
<td><p>Generic Network class.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="generated/pysegcnn.core.models.UNet.html#pysegcnn.core.models.UNet" title="pysegcnn.core.models.UNet"><code class="xref py py-obj docutils literal notranslate"><span class="pre">core.models.UNet</span></code></a></p></td>
-<td><p>A PyTorch implementation of <a class="reference external" href="https://arxiv.org/abs/1505.04597">U-Net</a>.</p></td>
+<td><p>A slightly modified implementation of <a class="reference external" href="https://arxiv.org/abs/1505.04597">U-Net</a> in PyTorch.</p></td>
</tr>
</tbody>
</table>
diff --git a/docs/_build/html/source/core.html b/docs/_build/html/source/core.html
deleted file mode 100644
index 18c74c4f435175b42a37e5266a4183e9d99e6034..0000000000000000000000000000000000000000
--- a/docs/_build/html/source/core.html
+++ /dev/null
@@ -1,6848 +0,0 @@
-
-
-<!DOCTYPE html>
-<html class="writer-html5" lang="en" >
-<head>
- <meta charset="utf-8">
-
- <meta name="viewport" content="width=device-width, initial-scale=1.0">
-
- <title>core package — PySegCNN documentation</title>
-
-
-
- <link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
- <link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
-
-
-
-
-
-
-
- <!--[if lt IE 9]>
- <script src="../_static/js/html5shiv.min.js"></script>
- <![endif]-->
-
-
- <script type="text/javascript" id="documentation_options" data-url_root="../" src="../_static/documentation_options.js"></script>
- <script src="../_static/jquery.js"></script>
- <script src="../_static/underscore.js"></script>
- <script src="../_static/doctools.js"></script>
- <script src="../_static/language_data.js"></script>
-
- <script type="text/javascript" src="../_static/js/theme.js"></script>
-
-
- <link rel="index" title="Index" href="../genindex.html" />
- <link rel="search" title="Search" href="../search.html" />
- <link rel="next" title="main package" href="main.html" />
- <link rel="prev" title="pysegcnn" href="modules.html" />
-</head>
-
-<body class="wy-body-for-nav">
-
-
- <div class="wy-grid-for-nav">
-
- <nav data-toggle="wy-nav-shift" class="wy-nav-side">
- <div class="wy-side-scroll">
- <div class="wy-side-nav-search" >
-
-
-
- <a href="../index.html" class="icon icon-home" alt="Documentation Home"> PySegCNN
-
-
-
- </a>
-
-
-
-
-
-
-
-<div role="search">
- <form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
- <input type="text" name="q" placeholder="Search docs" />
- <input type="hidden" name="check_keywords" value="yes" />
- <input type="hidden" name="area" value="default" />
- </form>
-</div>
-
-
- </div>
-
-
- <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
-
-
-
-
-
-
- <ul class="current">
-<li class="toctree-l1 current"><a class="reference internal" href="modules.html">pysegcnn</a><ul class="current">
-<li class="toctree-l2 current"><a class="current reference internal" href="#">core package</a><ul>
-<li class="toctree-l3"><a class="reference internal" href="#submodules">Submodules</a></li>
-<li class="toctree-l3"><a class="reference internal" href="#module-core.cli">core.cli module</a><ul>
-<li class="toctree-l4"><a class="reference internal" href="#license">License</a></li>
-</ul>
-</li>
-<li class="toctree-l3"><a class="reference internal" href="#module-core.constants">core.constants module</a><ul>
-<li class="toctree-l4"><a class="reference internal" href="#id1">License</a></li>
-</ul>
-</li>
-<li class="toctree-l3"><a class="reference internal" href="#module-core.dataset">core.dataset module</a><ul>
-<li class="toctree-l4"><a class="reference internal" href="#id2">License</a></li>
-</ul>
-</li>
-<li class="toctree-l3"><a class="reference internal" href="#module-core.graphics">core.graphics module</a><ul>
-<li class="toctree-l4"><a class="reference internal" href="#id4">License</a></li>
-</ul>
-</li>
-<li class="toctree-l3"><a class="reference internal" href="#module-core.layers">core.layers module</a><ul>
-<li class="toctree-l4"><a class="reference internal" href="#id5">License</a></li>
-</ul>
-</li>
-<li class="toctree-l3"><a class="reference internal" href="#module-core.logging">core.logging module</a><ul>
-<li class="toctree-l4"><a class="reference internal" href="#id6">License</a></li>
-</ul>
-</li>
-<li class="toctree-l3"><a class="reference internal" href="#module-core.models">core.models module</a><ul>
-<li class="toctree-l4"><a class="reference internal" href="#id7">License</a></li>
-</ul>
-</li>
-<li class="toctree-l3"><a class="reference internal" href="#module-core.predict">core.predict module</a><ul>
-<li class="toctree-l4"><a class="reference internal" href="#id8">License</a></li>
-</ul>
-</li>
-<li class="toctree-l3"><a class="reference internal" href="#module-core.split">core.split module</a><ul>
-<li class="toctree-l4"><a class="reference internal" href="#id9">License</a></li>
-</ul>
-</li>
-<li class="toctree-l3"><a class="reference internal" href="#module-core.trainer">core.trainer module</a><ul>
-<li class="toctree-l4"><a class="reference internal" href="#id10">License</a></li>
-</ul>
-</li>
-<li class="toctree-l3"><a class="reference internal" href="#module-core.transforms">core.transforms module</a><ul>
-<li class="toctree-l4"><a class="reference internal" href="#id11">License</a></li>
-</ul>
-</li>
-<li class="toctree-l3"><a class="reference internal" href="#module-core.utils">core.utils module</a><ul>
-<li class="toctree-l4"><a class="reference internal" href="#id12">License</a></li>
-</ul>
-</li>
-<li class="toctree-l3"><a class="reference internal" href="#module-core">Module contents</a></li>
-</ul>
-</li>
-<li class="toctree-l2"><a class="reference internal" href="main.html">main package</a></li>
-<li class="toctree-l2"><a class="reference internal" href="preprocessing.html">preprocessing package</a></li>
-</ul>
-</li>
-</ul>
-
-
-
- </div>
-
- </div>
- </nav>
-
- <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
-
-
- <nav class="wy-nav-top" aria-label="top navigation">
-
- <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
- <a href="../index.html">PySegCNN</a>
-
- </nav>
-
-
- <div class="wy-nav-content">
-
- <div class="rst-content">
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-<div role="navigation" aria-label="breadcrumbs navigation">
-
- <ul class="wy-breadcrumbs">
-
- <li><a href="../index.html" class="icon icon-home"></a> »</li>
-
- <li><a href="modules.html">pysegcnn</a> »</li>
-
- <li>core package</li>
-
-
- <li class="wy-breadcrumbs-aside">
-
-
- <a href="../_sources/source/core.rst.txt" rel="nofollow"> View page source</a>
-
-
- </li>
-
- </ul>
-
-
- <hr/>
-</div>
- <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
- <div itemprop="articleBody">
-
- <div class="section" id="core-package">
-<h1>core package<a class="headerlink" href="#core-package" title="Permalink to this headline">¶</a></h1>
-<div class="section" id="submodules">
-<h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this headline">¶</a></h2>
-</div>
-<div class="section" id="module-core.cli">
-<span id="core-cli-module"></span><h2>core.cli module<a class="headerlink" href="#module-core.cli" title="Permalink to this headline">¶</a></h2>
-<p>Command line interface parsers.</p>
-<div class="section" id="license">
-<h3>License<a class="headerlink" href="#license" title="Permalink to this headline">¶</a></h3>
-<blockquote>
-<div><p>Copyright (c) 2020 Daniel Frisinghelli</p>
-<p>This source code is licensed under the GNU General Public License v3.</p>
-<p>See the LICENSE file in the repository’s root directory.</p>
-</div></blockquote>
-<dl class="py function">
-<dt id="core.cli.structure_parser">
-<code class="sig-prename descclassname">core.cli.</code><code class="sig-name descname">structure_parser</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/cli.html#structure_parser"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.cli.structure_parser" title="Permalink to this definition">¶</a></dt>
-<dd><p>Command line argument parser to standardize dataset structure.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p></p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p>None.</p>
-</dd>
-</dl>
-</dd></dl>
-
-</div>
-</div>
-<div class="section" id="module-core.constants">
-<span id="core-constants-module"></span><h2>core.constants module<a class="headerlink" href="#module-core.constants" title="Permalink to this headline">¶</a></h2>
-<p>A collection of constant values.</p>
-<div class="section" id="id1">
-<h3>License<a class="headerlink" href="#id1" title="Permalink to this headline">¶</a></h3>
-<blockquote>
-<div><p>Copyright (c) 2020 Daniel Frisinghelli</p>
-<p>This source code is licensed under the GNU General Public License v3.</p>
-<p>See the LICENSE file in the repository’s root directory.</p>
-</div></blockquote>
-<dl class="py class">
-<dt id="core.constants.Cloud95Labels">
-<em class="property">class </em><code class="sig-prename descclassname">core.constants.</code><code class="sig-name descname">Cloud95Labels</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">value</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/constants.html#Cloud95Labels"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.constants.Cloud95Labels" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.constants.Label" title="core.constants.Label"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.constants.Label</span></code></a></p>
-<p>Class labels of the <a class="reference external" href="https://github.com/SorourMo/95-Cloud-An-Extension-to-38-Cloud-Dataset">Cloud-95</a> dataset.</p>
-<dl class="py attribute">
-<dt id="core.constants.Cloud95Labels.Clear">
-<code class="sig-name descname">Clear</code><em class="property"> = (0, 'skyblue')</em><a class="headerlink" href="#core.constants.Cloud95Labels.Clear" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.Cloud95Labels.Cloud">
-<code class="sig-name descname">Cloud</code><em class="property"> = (1, 'white')</em><a class="headerlink" href="#core.constants.Cloud95Labels.Cloud" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.constants.Label">
-<em class="property">class </em><code class="sig-prename descclassname">core.constants.</code><code class="sig-name descname">Label</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">value</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/constants.html#Label"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.constants.Label" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">enum.Enum</span></code></p>
-<p>Generic enumeration for class labels.</p>
-<dl class="py method">
-<dt id="core.constants.Label.color">
-<em class="property">property </em><code class="sig-name descname">color</code><a class="headerlink" href="#core.constants.Label.color" title="Permalink to this definition">¶</a></dt>
-<dd><p>Return the color to plot a class.</p>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.constants.Label.id">
-<em class="property">property </em><code class="sig-name descname">id</code><a class="headerlink" href="#core.constants.Label.id" title="Permalink to this definition">¶</a></dt>
-<dd><p>Return the value of a class in the ground truth.</p>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.constants.Landsat8">
-<em class="property">class </em><code class="sig-prename descclassname">core.constants.</code><code class="sig-name descname">Landsat8</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">value</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/constants.html#Landsat8"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.constants.Landsat8" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">enum.Enum</span></code></p>
-<p>The spectral bands of the Landsat 8 sensors.</p>
-<dl class="simple">
-<dt>sensors:</dt><dd><ul class="simple">
-<li><p>Operational Land Imager (OLI), (bands 1-9)</p></li>
-<li><p>Thermal Infrared Sensor (TIRS), (bands 10, 11)</p></li>
-</ul>
-</dd>
-</dl>
-<dl class="py attribute">
-<dt id="core.constants.Landsat8.blue">
-<code class="sig-name descname">blue</code><em class="property"> = 2</em><a class="headerlink" href="#core.constants.Landsat8.blue" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.Landsat8.cirrus">
-<code class="sig-name descname">cirrus</code><em class="property"> = 9</em><a class="headerlink" href="#core.constants.Landsat8.cirrus" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.Landsat8.green">
-<code class="sig-name descname">green</code><em class="property"> = 3</em><a class="headerlink" href="#core.constants.Landsat8.green" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.Landsat8.nir">
-<code class="sig-name descname">nir</code><em class="property"> = 5</em><a class="headerlink" href="#core.constants.Landsat8.nir" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.Landsat8.pan">
-<code class="sig-name descname">pan</code><em class="property"> = 8</em><a class="headerlink" href="#core.constants.Landsat8.pan" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.Landsat8.red">
-<code class="sig-name descname">red</code><em class="property"> = 4</em><a class="headerlink" href="#core.constants.Landsat8.red" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.Landsat8.swir1">
-<code class="sig-name descname">swir1</code><em class="property"> = 6</em><a class="headerlink" href="#core.constants.Landsat8.swir1" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.Landsat8.swir2">
-<code class="sig-name descname">swir2</code><em class="property"> = 7</em><a class="headerlink" href="#core.constants.Landsat8.swir2" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.Landsat8.tir1">
-<code class="sig-name descname">tir1</code><em class="property"> = 10</em><a class="headerlink" href="#core.constants.Landsat8.tir1" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.Landsat8.violet">
-<code class="sig-name descname">violet</code><em class="property"> = 1</em><a class="headerlink" href="#core.constants.Landsat8.violet" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.constants.ProSnowLabels">
-<em class="property">class </em><code class="sig-prename descclassname">core.constants.</code><code class="sig-name descname">ProSnowLabels</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">value</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/constants.html#ProSnowLabels"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.constants.ProSnowLabels" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.constants.Label" title="core.constants.Label"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.constants.Label</span></code></a></p>
-<p>Class labels of the ProSnow datasets.</p>
-<dl class="py attribute">
-<dt id="core.constants.ProSnowLabels.Cloud">
-<code class="sig-name descname">Cloud</code><em class="property"> = (0, 'white')</em><a class="headerlink" href="#core.constants.ProSnowLabels.Cloud" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.ProSnowLabels.Snow">
-<code class="sig-name descname">Snow</code><em class="property"> = (1, 'lightblue')</em><a class="headerlink" href="#core.constants.ProSnowLabels.Snow" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.ProSnowLabels.Snow_free">
-<code class="sig-name descname">Snow_free</code><em class="property"> = (2, 'sienna')</em><a class="headerlink" href="#core.constants.ProSnowLabels.Snow_free" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.constants.Sentinel2">
-<em class="property">class </em><code class="sig-prename descclassname">core.constants.</code><code class="sig-name descname">Sentinel2</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">value</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/constants.html#Sentinel2"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.constants.Sentinel2" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">enum.Enum</span></code></p>
-<p>The spectral bands of the Sentinel-2 MultiSpectral Instrument (MSI).</p>
-<dl class="py attribute">
-<dt id="core.constants.Sentinel2.aerosol">
-<code class="sig-name descname">aerosol</code><em class="property"> = 1</em><a class="headerlink" href="#core.constants.Sentinel2.aerosol" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.Sentinel2.blue">
-<code class="sig-name descname">blue</code><em class="property"> = 2</em><a class="headerlink" href="#core.constants.Sentinel2.blue" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.Sentinel2.cirrus">
-<code class="sig-name descname">cirrus</code><em class="property"> = 10</em><a class="headerlink" href="#core.constants.Sentinel2.cirrus" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.Sentinel2.green">
-<code class="sig-name descname">green</code><em class="property"> = 3</em><a class="headerlink" href="#core.constants.Sentinel2.green" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.Sentinel2.nir">
-<code class="sig-name descname">nir</code><em class="property"> = 8</em><a class="headerlink" href="#core.constants.Sentinel2.nir" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.Sentinel2.nnir">
-<code class="sig-name descname">nnir</code><em class="property"> = '8A'</em><a class="headerlink" href="#core.constants.Sentinel2.nnir" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.Sentinel2.red">
-<code class="sig-name descname">red</code><em class="property"> = 4</em><a class="headerlink" href="#core.constants.Sentinel2.red" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.Sentinel2.swir1">
-<code class="sig-name descname">swir1</code><em class="property"> = 11</em><a class="headerlink" href="#core.constants.Sentinel2.swir1" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.Sentinel2.swir2">
-<code class="sig-name descname">swir2</code><em class="property"> = 12</em><a class="headerlink" href="#core.constants.Sentinel2.swir2" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.Sentinel2.vapor">
-<code class="sig-name descname">vapor</code><em class="property"> = 9</em><a class="headerlink" href="#core.constants.Sentinel2.vapor" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.Sentinel2.vnir1">
-<code class="sig-name descname">vnir1</code><em class="property"> = 5</em><a class="headerlink" href="#core.constants.Sentinel2.vnir1" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.Sentinel2.vnir2">
-<code class="sig-name descname">vnir2</code><em class="property"> = 6</em><a class="headerlink" href="#core.constants.Sentinel2.vnir2" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.Sentinel2.vnir3">
-<code class="sig-name descname">vnir3</code><em class="property"> = 7</em><a class="headerlink" href="#core.constants.Sentinel2.vnir3" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.constants.SparcsLabels">
-<em class="property">class </em><code class="sig-prename descclassname">core.constants.</code><code class="sig-name descname">SparcsLabels</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">value</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/constants.html#SparcsLabels"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.constants.SparcsLabels" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.constants.Label" title="core.constants.Label"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.constants.Label</span></code></a></p>
-<p>Class labels of the <a class="reference external" href="https://www.usgs.gov/land-resources/nli/landsat/spatial-procedures-automated-removal-cloud-and-shadow-sparcs-validation">Sparcs</a> dataset.</p>
-<dl class="py attribute">
-<dt id="core.constants.SparcsLabels.Cloud">
-<code class="sig-name descname">Cloud</code><em class="property"> = (5, 'white')</em><a class="headerlink" href="#core.constants.SparcsLabels.Cloud" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.SparcsLabels.Flooded">
-<code class="sig-name descname">Flooded</code><em class="property"> = (6, 'yellow')</em><a class="headerlink" href="#core.constants.SparcsLabels.Flooded" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.SparcsLabels.Land">
-<code class="sig-name descname">Land</code><em class="property"> = (4, 'sienna')</em><a class="headerlink" href="#core.constants.SparcsLabels.Land" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.SparcsLabels.Shadow">
-<code class="sig-name descname">Shadow</code><em class="property"> = (0, 'grey')</em><a class="headerlink" href="#core.constants.SparcsLabels.Shadow" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.SparcsLabels.Shadow_over_water">
-<code class="sig-name descname">Shadow_over_water</code><em class="property"> = (1, 'darkblue')</em><a class="headerlink" href="#core.constants.SparcsLabels.Shadow_over_water" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.SparcsLabels.Snow">
-<code class="sig-name descname">Snow</code><em class="property"> = (3, 'lightblue')</em><a class="headerlink" href="#core.constants.SparcsLabels.Snow" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.constants.SparcsLabels.Water">
-<code class="sig-name descname">Water</code><em class="property"> = (2, 'blue')</em><a class="headerlink" href="#core.constants.SparcsLabels.Water" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-</dd></dl>
-
-</div>
-</div>
-<div class="section" id="module-core.dataset">
-<span id="core-dataset-module"></span><h2>core.dataset module<a class="headerlink" href="#module-core.dataset" title="Permalink to this headline">¶</a></h2>
-<p>Custom dataset classes compliant to the PyTorch standard.</p>
-<p>Each custom dataset should inherit from torch.utils.data.Dataset to benefit
-from the torch.utils.data.DataLoader class, which implements helpful utilities
-during model training.</p>
-<p>For any kind of image-like dataset, inherit the ImageDataset class to create
-your custom dataset.</p>
-<div class="section" id="id2">
-<h3>License<a class="headerlink" href="#id2" title="Permalink to this headline">¶</a></h3>
-<blockquote>
-<div><p>Copyright (c) 2020 Daniel Frisinghelli</p>
-<p>This source code is licensed under the GNU General Public License v3.</p>
-<p>See the LICENSE file in the repository’s root directory.</p>
-</div></blockquote>
-<dl class="py class">
-<dt id="core.dataset.Cloud95Dataset">
-<em class="property">class </em><code class="sig-prename descclassname">core.dataset.</code><code class="sig-name descname">Cloud95Dataset</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">root_dir</span></em>, <em class="sig-param"><span class="n">use_bands</span><span class="o">=</span><span class="default_value">[]</span></em>, <em class="sig-param"><span class="n">tile_size</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="n">pad</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">gt_pattern</span><span class="o">=</span><span class="default_value">'(.*)gt\\.tif'</span></em>, <em class="sig-param"><span class="n">sort</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">seed</span><span class="o">=</span><span class="default_value">0</span></em>, <em class="sig-param"><span class="n">transforms</span><span class="o">=</span><span class="default_value">[]</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#Cloud95Dataset"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.Cloud95Dataset" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.dataset.ImageDataset" title="core.dataset.ImageDataset"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.dataset.ImageDataset</span></code></a></p>
-<p>Class for the <a class="reference external" href="https://github.com/SorourMo/95-Cloud-An-Extension-to-38-Cloud-Dataset">Cloud-95</a> dataset by <a class="reference external" href="https://arxiv.org/abs/2001.08768">Mohajerani et al. (2020)</a>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>root_dir</strong> (<cite>str</cite>) – The root directory, path to the dataset.</p></li>
-<li><p><strong>use_bands</strong> (<cite>list</cite> [<cite>str</cite>], optional) – A list of the spectral bands to use. The default is [].</p></li>
-<li><p><strong>tile_size</strong> (<cite>int</cite> or <cite>None</cite>, optional) – The size of the tiles. If not <cite>None</cite>, each scene is divided into square
-tiles of shape (tile_size, tile_size). The default is None.</p></li>
-<li><p><strong>pad</strong> (<cite>bool</cite>, optional) – Whether to center pad the input image. Set <code class="docutils literal notranslate"><span class="pre">pad</span></code> = True, if the
-images are not evenly divisible by the <code class="docutils literal notranslate"><span class="pre">tile_size</span></code>. The image data is
-padded with a constant padding value of zero. For each image, the
-corresponding ground truth image is padded with a “no data†label.
-The default is False.</p></li>
-<li><p><strong>gt_pattern</strong> (<cite>str</cite>, optional) – A regural expression to match the ground truth naming convention. All
-directories and subdirectories in <code class="docutils literal notranslate"><span class="pre">root_dir</span></code> are searched for files
-matching <code class="docutils literal notranslate"><span class="pre">gt_pattern</span></code>. The default is ‘(.*)gt\.tif’.</p></li>
-<li><p><strong>sort</strong> (<cite>bool</cite>, optional) – Whether to chronologically sort the samples. Useful for time series
-data. The default is False.</p></li>
-<li><p><strong>seed</strong> (<cite>int</cite>, optional) – The random seed. Used to split the dataset into training, validation
-and test set. Useful for reproducibility. The default is 0.</p></li>
-<li><p><strong>transforms</strong> (<cite>list</cite> [<cite>pysegcnn.core.split.Augment</cite>], optional) – List of <cite>pysegcnn.core.split.Augment</cite> instances. Each item in
-<code class="docutils literal notranslate"><span class="pre">transforms</span></code> generates a distinct transformed version of the dataset.
-The total dataset is composed of the original untransformed dataset
-together with each transformed version of it.
-If <code class="docutils literal notranslate"><span class="pre">transforms</span></code> = [], only the original dataset is used.
-The default is [].</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">build_samples</span></code>(scene)</p></td>
-<td><p>Stack the bands of a sample in a single array.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.dataset.Cloud95Dataset.compose_scenes" title="core.dataset.Cloud95Dataset.compose_scenes"><code class="xref py py-obj docutils literal notranslate"><span class="pre">compose_scenes</span></code></a>()</p></td>
-<td><p>Build the list of samples of the dataset.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.dataset.Cloud95Dataset.get_labels" title="core.dataset.Cloud95Dataset.get_labels"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_labels</span></code></a>()</p></td>
-<td><p>Class labels of the Cloud-95 dataset.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.dataset.Cloud95Dataset.get_sensor" title="core.dataset.Cloud95Dataset.get_sensor"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_sensor</span></code></a>()</p></td>
-<td><p>Landsat 8 bands of the Cloud-95 dataset.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.dataset.Cloud95Dataset.get_size" title="core.dataset.Cloud95Dataset.get_size"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_size</span></code></a>()</p></td>
-<td><p>Image size of the Cloud-95 dataset.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.dataset.Cloud95Dataset.parse_scene_id" title="core.dataset.Cloud95Dataset.parse_scene_id"><code class="xref py py-obj docutils literal notranslate"><span class="pre">parse_scene_id</span></code></a>(scene_id)</p></td>
-<td><p>Parse Sparcs scene identifiers (Landsat 8).</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.dataset.Cloud95Dataset.preprocess" title="core.dataset.Cloud95Dataset.preprocess"><code class="xref py py-obj docutils literal notranslate"><span class="pre">preprocess</span></code></a>(data, gt)</p></td>
-<td><p>Preprocess Cloud-95 dataset images.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">read_scene</span></code>(idx)</p></td>
-<td><p>Read the data of the sample with index <code class="docutils literal notranslate"><span class="pre">idx</span></code>.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to_tensor</span></code>(x, dtype)</p></td>
-<td><p>Convert <code class="docutils literal notranslate"><span class="pre">x</span></code> to <cite>torch.Tensor</cite>.</p></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="core.dataset.Cloud95Dataset.compose_scenes">
-<code class="sig-name descname">compose_scenes</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#Cloud95Dataset.compose_scenes"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.Cloud95Dataset.compose_scenes" title="Permalink to this definition">¶</a></dt>
-<dd><p>Build the list of samples of the dataset.</p>
-<p>Each sample is represented by a dictionary.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><p><strong>scenes</strong> –</p>
-<dl class="simple">
-<dt>Each item in <code class="docutils literal notranslate"><span class="pre">scenes</span></code> is a <cite>dict</cite> with keys:</dt><dd><dl class="simple">
-<dt><code class="docutils literal notranslate"><span class="pre">'band_name_1'</span></code></dt><dd><p>Path to the file of band_1.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'band_name_2'</span></code></dt><dd><p>Path to the file of band_2.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'band_name_n'</span></code></dt><dd><p>Path to the file of band_n.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'gt'</span></code></dt><dd><p>Path to the ground truth file.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'date'</span></code></dt><dd><p>The date of the sample.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'tile'</span></code></dt><dd><p>The tile id of the sample.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'transform'</span></code></dt><dd><p>The transformation to apply.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'id'</span></code></dt><dd><p>The scene identifier.</p>
-</dd>
-</dl>
-</dd>
-</dl>
-</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>list</cite> [<cite>dict</cite>]</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.dataset.Cloud95Dataset.get_labels">
-<code class="sig-name descname">get_labels</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#Cloud95Dataset.get_labels"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.Cloud95Dataset.get_labels" title="Permalink to this definition">¶</a></dt>
-<dd><p>Class labels of the Cloud-95 dataset.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>labels</strong> – The class labels.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>enum.Enum</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.dataset.Cloud95Dataset.get_sensor">
-<code class="sig-name descname">get_sensor</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#Cloud95Dataset.get_sensor"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.Cloud95Dataset.get_sensor" title="Permalink to this definition">¶</a></dt>
-<dd><p>Landsat 8 bands of the Cloud-95 dataset.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>sensor</strong> – An enumeration of the bands of the sensor.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>enum.Enum</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.dataset.Cloud95Dataset.get_size">
-<code class="sig-name descname">get_size</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#Cloud95Dataset.get_size"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.Cloud95Dataset.get_size" title="Permalink to this definition">¶</a></dt>
-<dd><p>Image size of the Cloud-95 dataset.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>size</strong> – The image size (height, width).</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>tuple</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.dataset.Cloud95Dataset.parse_scene_id">
-<code class="sig-name descname">parse_scene_id</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">scene_id</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#Cloud95Dataset.parse_scene_id"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.Cloud95Dataset.parse_scene_id" title="Permalink to this definition">¶</a></dt>
-<dd><p>Parse Sparcs scene identifiers (Landsat 8).</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>scene_id</strong> (<cite>str</cite>) – A scene identifier.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>scene</strong> – A dictionary containing scene metadata. If <cite>None</cite>, <code class="docutils literal notranslate"><span class="pre">scene_id</span></code> is
-not a valid Landsat scene identifier.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>dict</cite> or <cite>None</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.dataset.Cloud95Dataset.preprocess">
-<code class="sig-name descname">preprocess</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">data</span></em>, <em class="sig-param"><span class="n">gt</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#Cloud95Dataset.preprocess"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.Cloud95Dataset.preprocess" title="Permalink to this definition">¶</a></dt>
-<dd><p>Preprocess Cloud-95 dataset images.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>data</strong> (<cite>numpy.ndarray</cite>) – The sample input data.</p></li>
-<li><p><strong>gt</strong> (<cite>numpy.ndarray</cite>) – The sample ground truth.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><ul class="simple">
-<li><p><strong>data</strong> (<cite>numpy.ndarray</cite>) – The preprocessed input data.</p></li>
-<li><p><strong>gt</strong> (<cite>numpy.ndarray</cite>) – The preprocessed ground truth data.</p></li>
-</ul>
-</p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.dataset.ImageDataset">
-<em class="property">class </em><code class="sig-prename descclassname">core.dataset.</code><code class="sig-name descname">ImageDataset</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">root_dir</span></em>, <em class="sig-param"><span class="n">use_bands</span><span class="o">=</span><span class="default_value">[]</span></em>, <em class="sig-param"><span class="n">tile_size</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="n">pad</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">gt_pattern</span><span class="o">=</span><span class="default_value">'(.*)gt.tif'</span></em>, <em class="sig-param"><span class="n">sort</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">seed</span><span class="o">=</span><span class="default_value">0</span></em>, <em class="sig-param"><span class="n">transforms</span><span class="o">=</span><span class="default_value">[]</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#ImageDataset"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.ImageDataset" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">torch.utils.data.dataset.Dataset</span></code></p>
-<p>Base class for multispectral image data.</p>
-<p>Inheriting from <cite>torch.utils.data.Dataset</cite> to be compliant to the PyTorch
-standard. Furthermore, using instances of <cite>torch.utils.data.Dataset</cite>
-enables the use of the handy <cite>torch.utils.data.DataLoader</cite> class during
-model training.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>root_dir</strong> (<cite>str</cite>) – The root directory, path to the dataset.</p></li>
-<li><p><strong>use_bands</strong> (<cite>list</cite> [<cite>str</cite>], optional) – A list of the spectral bands to use. The default is [].</p></li>
-<li><p><strong>tile_size</strong> (<cite>int</cite> or <cite>None</cite>, optional) – The size of the tiles. If not <cite>None</cite>, each scene is divided into square
-tiles of shape (tile_size, tile_size). The default is None.</p></li>
-<li><p><strong>pad</strong> (<cite>bool</cite>, optional) – Whether to center pad the input image. Set <code class="docutils literal notranslate"><span class="pre">pad</span></code> = True, if the
-images are not evenly divisible by the <code class="docutils literal notranslate"><span class="pre">tile_size</span></code>. The image data is
-padded with a constant padding value of zero. For each image, the
-corresponding ground truth image is padded with a “no data†label.
-The default is False.</p></li>
-<li><p><strong>gt_pattern</strong> (<cite>str</cite>, optional) – A regural expression to match the ground truth naming convention.
-All directories and subdirectories in <code class="docutils literal notranslate"><span class="pre">root_dir</span></code> are searched for
-files matching <code class="docutils literal notranslate"><span class="pre">gt_pattern</span></code>. The default is ‘(.*)gt\.tif’.</p></li>
-<li><p><strong>sort</strong> (<cite>bool</cite>, optional) – Whether to chronologically sort the samples. Useful for time series
-data. The default is False.</p></li>
-<li><p><strong>seed</strong> (<cite>int</cite>, optional) – The random seed. Used to split the dataset into training, validation
-and test set. Useful for reproducibility. The default is 0.</p></li>
-<li><p><strong>transforms</strong> (<cite>list</cite> [<cite>pysegcnn.core.split.Augment</cite>], optional) – List of <cite>pysegcnn.core.split.Augment</cite> instances. Each item in
-<code class="docutils literal notranslate"><span class="pre">transforms</span></code> generates a distinct transformed version of the dataset.
-The total dataset is composed of the original untransformed dataset
-together with each transformed version of it.
-If <code class="docutils literal notranslate"><span class="pre">transforms</span></code> = [], only the original dataset is used.
-The default is [].</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.dataset.ImageDataset.build_samples" title="core.dataset.ImageDataset.build_samples"><code class="xref py py-obj docutils literal notranslate"><span class="pre">build_samples</span></code></a>(scene)</p></td>
-<td><p>Stack the bands of a sample in a single array.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.dataset.ImageDataset.compose_scenes" title="core.dataset.ImageDataset.compose_scenes"><code class="xref py py-obj docutils literal notranslate"><span class="pre">compose_scenes</span></code></a>()</p></td>
-<td><p>Build the list of samples of the dataset.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.dataset.ImageDataset.get_labels" title="core.dataset.ImageDataset.get_labels"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_labels</span></code></a>()</p></td>
-<td><p>Return an enumeration of the class labels of the dataset.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.dataset.ImageDataset.get_sensor" title="core.dataset.ImageDataset.get_sensor"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_sensor</span></code></a>()</p></td>
-<td><p>Return an enumeration of the bands of the sensor of the dataset.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.dataset.ImageDataset.get_size" title="core.dataset.ImageDataset.get_size"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_size</span></code></a>()</p></td>
-<td><p>Return the size of the images in the dataset.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.dataset.ImageDataset.parse_scene_id" title="core.dataset.ImageDataset.parse_scene_id"><code class="xref py py-obj docutils literal notranslate"><span class="pre">parse_scene_id</span></code></a>(scene_id)</p></td>
-<td><p>Parse the scene identifier.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.dataset.ImageDataset.preprocess" title="core.dataset.ImageDataset.preprocess"><code class="xref py py-obj docutils literal notranslate"><span class="pre">preprocess</span></code></a>(data, gt)</p></td>
-<td><p>Preprocess a sample before feeding it to a model.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.dataset.ImageDataset.read_scene" title="core.dataset.ImageDataset.read_scene"><code class="xref py py-obj docutils literal notranslate"><span class="pre">read_scene</span></code></a>(idx)</p></td>
-<td><p>Read the data of the sample with index <code class="docutils literal notranslate"><span class="pre">idx</span></code>.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.dataset.ImageDataset.to_tensor" title="core.dataset.ImageDataset.to_tensor"><code class="xref py py-obj docutils literal notranslate"><span class="pre">to_tensor</span></code></a>(x, dtype)</p></td>
-<td><p>Convert <code class="docutils literal notranslate"><span class="pre">x</span></code> to <cite>torch.Tensor</cite>.</p></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="core.dataset.ImageDataset.build_samples">
-<code class="sig-name descname">build_samples</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">scene</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#ImageDataset.build_samples"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.ImageDataset.build_samples" title="Permalink to this definition">¶</a></dt>
-<dd><p>Stack the bands of a sample in a single array.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>scene</strong> (<cite>dict</cite>) – <dl class="simple">
-<dt>The sample data dictionary with keys:</dt><dd><dl class="simple">
-<dt><code class="docutils literal notranslate"><span class="pre">'band_name_1'</span></code></dt><dd><p>data of band_1 (<cite>numpy.ndarray</cite>).</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'band_name_2'</span></code></dt><dd><p>data of band_2 (<cite>numpy.ndarray</cite>).</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'band_name_n'</span></code></dt><dd><p>data of band_n (<cite>numpy.ndarray</cite>).</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'gt'</span></code></dt><dd><p>data of the ground truth (<cite>numpy.ndarray</cite>).</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'date'</span></code></dt><dd><p>The date of the sample.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'tile'</span></code></dt><dd><p>The tile id of the sample.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'transform'</span></code></dt><dd><p>The transformation to apply.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'id'</span></code></dt><dd><p>The scene identifier.</p>
-</dd>
-</dl>
-</dd>
-</dl>
-</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><ul class="simple">
-<li><p><strong>stack</strong> (<cite>numpy.ndarray</cite>) – The input data of the sample.</p></li>
-<li><p><strong>gt</strong> (<em>TYPE</em>) – The ground truth of the sample.</p></li>
-</ul>
-</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.dataset.ImageDataset.compose_scenes">
-<code class="sig-name descname">compose_scenes</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#ImageDataset.compose_scenes"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.ImageDataset.compose_scenes" title="Permalink to this definition">¶</a></dt>
-<dd><p>Build the list of samples of the dataset.</p>
-<p>Each sample is represented by a dictionary.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Raises</dt>
-<dd class="field-odd"><p><strong>NotImplementedError</strong> – Raised if the <cite>pysegcnn.core.dataset.ImageDataset</cite> class is not
-inherited.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><p><strong>samples</strong> –</p>
-<dl class="simple">
-<dt>Each dictionary representing a sample should have keys:</dt><dd><dl class="simple">
-<dt><code class="docutils literal notranslate"><span class="pre">'band_name_1'</span></code></dt><dd><p>Path to the file of band_1.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'band_name_2'</span></code></dt><dd><p>Path to the file of band_2.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'band_name_n'</span></code></dt><dd><p>Path to the file of band_n.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'gt'</span></code></dt><dd><p>Path to the ground truth file.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'date'</span></code></dt><dd><p>The date of the sample.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'tile'</span></code></dt><dd><p>The tile id of the sample.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'transform'</span></code></dt><dd><p>The transformation to apply.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'id'</span></code></dt><dd><p>The scene identifier.</p>
-</dd>
-</dl>
-</dd>
-</dl>
-</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>list</cite> [<cite>dict</cite>]</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.dataset.ImageDataset.get_labels">
-<code class="sig-name descname">get_labels</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#ImageDataset.get_labels"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.ImageDataset.get_labels" title="Permalink to this definition">¶</a></dt>
-<dd><p>Return an enumeration of the class labels of the dataset.</p>
-<p>Examples can be found in <cite>pysegcnn.core.constants</cite>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Raises</dt>
-<dd class="field-odd"><p><strong>NotImplementedError</strong> – Raised if the <cite>pysegcnn.core.dataset.ImageDataset</cite> class is not
-inherited.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>labels</strong> – The class labels.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>enum.Enum</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.dataset.ImageDataset.get_sensor">
-<code class="sig-name descname">get_sensor</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#ImageDataset.get_sensor"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.ImageDataset.get_sensor" title="Permalink to this definition">¶</a></dt>
-<dd><p>Return an enumeration of the bands of the sensor of the dataset.</p>
-<p>Examples can be found in <cite>pysegcnn.core.constants</cite>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Raises</dt>
-<dd class="field-odd"><p><strong>NotImplementedError</strong> – Raised if the <cite>pysegcnn.core.dataset.ImageDataset</cite> class is not
-inherited.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>sensor</strong> – An enumeration of the bands of the sensor.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>enum.Enum</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.dataset.ImageDataset.get_size">
-<code class="sig-name descname">get_size</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#ImageDataset.get_size"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.ImageDataset.get_size" title="Permalink to this definition">¶</a></dt>
-<dd><p>Return the size of the images in the dataset.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Raises</dt>
-<dd class="field-odd"><p><strong>NotImplementedError</strong> – Raised if the <cite>pysegcnn.core.dataset.ImageDataset</cite> class is not
-inherited.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>size</strong> – The image size (height, width).</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>tuple</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.dataset.ImageDataset.parse_scene_id">
-<code class="sig-name descname">parse_scene_id</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">scene_id</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#ImageDataset.parse_scene_id"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.ImageDataset.parse_scene_id" title="Permalink to this definition">¶</a></dt>
-<dd><p>Parse the scene identifier.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>scene_id</strong> (<cite>str</cite>) – A scene identifier.</p>
-</dd>
-<dt class="field-even">Raises</dt>
-<dd class="field-even"><p><strong>NotImplementedError</strong> – Raised if the <cite>pysegcnn.core.dataset.ImageDataset</cite> class is not
-inherited.</p>
-</dd>
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>scene</strong> – A dictionary containing scene metadata. If <cite>None</cite>, <code class="docutils literal notranslate"><span class="pre">scene_id</span></code> is
-not a valid scene identifier.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>dict</cite> or <cite>None</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.dataset.ImageDataset.preprocess">
-<code class="sig-name descname">preprocess</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">data</span></em>, <em class="sig-param"><span class="n">gt</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#ImageDataset.preprocess"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.ImageDataset.preprocess" title="Permalink to this definition">¶</a></dt>
-<dd><p>Preprocess a sample before feeding it to a model.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>data</strong> (<cite>numpy.ndarray</cite>) – The sample input data.</p></li>
-<li><p><strong>gt</strong> (<cite>numpy.ndarray</cite>) – The sample ground truth.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Raises</dt>
-<dd class="field-even"><p><strong>NotImplementedError</strong> – Raised if the <cite>pysegcnn.core.dataset.ImageDataset</cite> class is not
-inherited.</p>
-</dd>
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><ul class="simple">
-<li><p><strong>data</strong> (<cite>numpy.ndarray</cite>) – The preprocessed input data.</p></li>
-<li><p><strong>gt</strong> (<cite>numpy.ndarray</cite>) – The preprocessed ground truth data.</p></li>
-</ul>
-</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.dataset.ImageDataset.read_scene">
-<code class="sig-name descname">read_scene</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">idx</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#ImageDataset.read_scene"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.ImageDataset.read_scene" title="Permalink to this definition">¶</a></dt>
-<dd><p>Read the data of the sample with index <code class="docutils literal notranslate"><span class="pre">idx</span></code>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>idx</strong> (<cite>int</cite>) – The index of the sample.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><p><strong>scene_data</strong> –</p>
-<dl class="simple">
-<dt>The sample data dictionary with keys:</dt><dd><dl class="simple">
-<dt><code class="docutils literal notranslate"><span class="pre">'band_name_1'</span></code></dt><dd><p>data of band_1 (<cite>numpy.ndarray</cite>).</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'band_name_2'</span></code></dt><dd><p>data of band_2 (<cite>numpy.ndarray</cite>).</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'band_name_n'</span></code></dt><dd><p>data of band_n (<cite>numpy.ndarray</cite>).</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'gt'</span></code></dt><dd><p>data of the ground truth (<cite>numpy.ndarray</cite>).</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'date'</span></code></dt><dd><p>The date of the sample.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'tile'</span></code></dt><dd><p>The tile id of the sample.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'transform'</span></code></dt><dd><p>The transformation to apply.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'id'</span></code></dt><dd><p>The scene identifier.</p>
-</dd>
-</dl>
-</dd>
-</dl>
-</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>dict</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.dataset.ImageDataset.to_tensor">
-<code class="sig-name descname">to_tensor</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">x</span></em>, <em class="sig-param"><span class="n">dtype</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#ImageDataset.to_tensor"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.ImageDataset.to_tensor" title="Permalink to this definition">¶</a></dt>
-<dd><p>Convert <code class="docutils literal notranslate"><span class="pre">x</span></code> to <cite>torch.Tensor</cite>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>x</strong> (<em>array_like</em>) – The input data.</p></li>
-<li><p><strong>dtype</strong> (<cite>torch.dtype</cite>) – The data type used to convert <code class="docutils literal notranslate"><span class="pre">x</span></code>.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>x</strong> – The input data tensor.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>torch.Tensor</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.dataset.ProSnowDataset">
-<em class="property">class </em><code class="sig-prename descclassname">core.dataset.</code><code class="sig-name descname">ProSnowDataset</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">root_dir</span></em>, <em class="sig-param"><span class="n">use_bands</span><span class="o">=</span><span class="default_value">[]</span></em>, <em class="sig-param"><span class="n">tile_size</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="n">pad</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">gt_pattern</span><span class="o">=</span><span class="default_value">'(.*)gt\\.tif'</span></em>, <em class="sig-param"><span class="n">sort</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">seed</span><span class="o">=</span><span class="default_value">0</span></em>, <em class="sig-param"><span class="n">transforms</span><span class="o">=</span><span class="default_value">[]</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#ProSnowDataset"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.ProSnowDataset" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.dataset.StandardEoDataset" title="core.dataset.StandardEoDataset"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.dataset.StandardEoDataset</span></code></a></p>
-<p>Class for the ProSnow datasets.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>root_dir</strong> (<cite>str</cite>) – The root directory, path to the dataset.</p></li>
-<li><p><strong>use_bands</strong> (<cite>list</cite> [<cite>str</cite>], optional) – A list of the spectral bands to use. The default is [].</p></li>
-<li><p><strong>tile_size</strong> (<cite>int</cite> or <cite>None</cite>, optional) – The size of the tiles. If not <cite>None</cite>, each scene is divided into square
-tiles of shape (tile_size, tile_size). The default is None.</p></li>
-<li><p><strong>pad</strong> (<cite>bool</cite>, optional) – Whether to center pad the input image. Set <code class="docutils literal notranslate"><span class="pre">pad</span></code> = True, if the
-images are not evenly divisible by the <code class="docutils literal notranslate"><span class="pre">tile_size</span></code>. The image data is
-padded with a constant padding value of zero. For each image, the
-corresponding ground truth image is padded with a “no data†label.
-The default is False.</p></li>
-<li><p><strong>gt_pattern</strong> (<cite>str</cite>, optional) – A regural expression to match the ground truth naming convention. All
-directories and subdirectories in <code class="docutils literal notranslate"><span class="pre">root_dir</span></code> are searched for files
-matching <code class="docutils literal notranslate"><span class="pre">gt_pattern</span></code>. The default is ‘(.*)gt\.tif’.</p></li>
-<li><p><strong>sort</strong> (<cite>bool</cite>, optional) – Whether to chronologically sort the samples. Useful for time series
-data. The default is False.</p></li>
-<li><p><strong>seed</strong> (<cite>int</cite>, optional) – The random seed. Used to split the dataset into training, validation
-and test set. Useful for reproducibility. The default is 0.</p></li>
-<li><p><strong>transforms</strong> (<cite>list</cite> [<cite>pysegcnn.core.split.Augment</cite>], optional) – List of <cite>pysegcnn.core.split.Augment</cite> instances. Each item in
-<code class="docutils literal notranslate"><span class="pre">transforms</span></code> generates a distinct transformed version of the dataset.
-The total dataset is composed of the original untransformed dataset
-together with each transformed version of it.
-If <code class="docutils literal notranslate"><span class="pre">transforms</span></code> = [], only the original dataset is used.
-The default is [].</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">build_samples</span></code>(scene)</p></td>
-<td><p>Stack the bands of a sample in a single array.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">compose_scenes</span></code>()</p></td>
-<td><p>Build the list of samples of the dataset.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.dataset.ProSnowDataset.get_labels" title="core.dataset.ProSnowDataset.get_labels"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_labels</span></code></a>()</p></td>
-<td><p>Class labels of the ProSnow datasets.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.dataset.ProSnowDataset.get_sensor" title="core.dataset.ProSnowDataset.get_sensor"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_sensor</span></code></a>()</p></td>
-<td><p>Sentinel 2 bands of the ProSnow datasets.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_size</span></code>()</p></td>
-<td><p>Return the size of the images in the dataset.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.dataset.ProSnowDataset.parse_scene_id" title="core.dataset.ProSnowDataset.parse_scene_id"><code class="xref py py-obj docutils literal notranslate"><span class="pre">parse_scene_id</span></code></a>(scene_id)</p></td>
-<td><p>Parse ProSnow scene identifiers (Sentinel 2).</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.dataset.ProSnowDataset.preprocess" title="core.dataset.ProSnowDataset.preprocess"><code class="xref py py-obj docutils literal notranslate"><span class="pre">preprocess</span></code></a>(data, gt)</p></td>
-<td><p>Preprocess ProSnow dataset images.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">read_scene</span></code>(idx)</p></td>
-<td><p>Read the data of the sample with index <code class="docutils literal notranslate"><span class="pre">idx</span></code>.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to_tensor</span></code>(x, dtype)</p></td>
-<td><p>Convert <code class="docutils literal notranslate"><span class="pre">x</span></code> to <cite>torch.Tensor</cite>.</p></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="core.dataset.ProSnowDataset.get_labels">
-<code class="sig-name descname">get_labels</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#ProSnowDataset.get_labels"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.ProSnowDataset.get_labels" title="Permalink to this definition">¶</a></dt>
-<dd><p>Class labels of the ProSnow datasets.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>labels</strong> – The class labels.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>enum.Enum</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.dataset.ProSnowDataset.get_sensor">
-<code class="sig-name descname">get_sensor</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#ProSnowDataset.get_sensor"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.ProSnowDataset.get_sensor" title="Permalink to this definition">¶</a></dt>
-<dd><p>Sentinel 2 bands of the ProSnow datasets.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>sensor</strong> – An enumeration of the bands of the sensor.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>enum.Enum</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.dataset.ProSnowDataset.parse_scene_id">
-<code class="sig-name descname">parse_scene_id</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">scene_id</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#ProSnowDataset.parse_scene_id"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.ProSnowDataset.parse_scene_id" title="Permalink to this definition">¶</a></dt>
-<dd><p>Parse ProSnow scene identifiers (Sentinel 2).</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>scene_id</strong> (<cite>str</cite>) – A scene identifier.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>scene</strong> – A dictionary containing scene metadata. If <cite>None</cite>, <code class="docutils literal notranslate"><span class="pre">scene_id</span></code> is
-not a valid Sentinel-2 scene identifier.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>dict</cite> or <cite>None</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.dataset.ProSnowDataset.preprocess">
-<code class="sig-name descname">preprocess</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">data</span></em>, <em class="sig-param"><span class="n">gt</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#ProSnowDataset.preprocess"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.ProSnowDataset.preprocess" title="Permalink to this definition">¶</a></dt>
-<dd><p>Preprocess ProSnow dataset images.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>data</strong> (<cite>numpy.ndarray</cite>) – The sample input data.</p></li>
-<li><p><strong>gt</strong> (<cite>numpy.ndarray</cite>) – The sample ground truth.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><ul class="simple">
-<li><p><strong>data</strong> (<cite>numpy.ndarray</cite>) – The preprocessed input data.</p></li>
-<li><p><strong>gt</strong> (<cite>numpy.ndarray</cite>) – The preprocessed ground truth data.</p></li>
-</ul>
-</p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.dataset.ProSnowGarmisch">
-<em class="property">class </em><code class="sig-prename descclassname">core.dataset.</code><code class="sig-name descname">ProSnowGarmisch</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">root_dir</span></em>, <em class="sig-param"><span class="n">use_bands</span><span class="o">=</span><span class="default_value">[]</span></em>, <em class="sig-param"><span class="n">tile_size</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="n">pad</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">gt_pattern</span><span class="o">=</span><span class="default_value">'(.*)gt\\.tif'</span></em>, <em class="sig-param"><span class="n">sort</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">seed</span><span class="o">=</span><span class="default_value">0</span></em>, <em class="sig-param"><span class="n">transforms</span><span class="o">=</span><span class="default_value">[]</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#ProSnowGarmisch"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.ProSnowGarmisch" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.dataset.ProSnowDataset" title="core.dataset.ProSnowDataset"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.dataset.ProSnowDataset</span></code></a></p>
-<p>Class for the ProSnow Garmisch dataset.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>root_dir</strong> (<cite>str</cite>) – The root directory, path to the dataset.</p></li>
-<li><p><strong>use_bands</strong> (<cite>list</cite> [<cite>str</cite>], optional) – A list of the spectral bands to use. The default is [].</p></li>
-<li><p><strong>tile_size</strong> (<cite>int</cite> or <cite>None</cite>, optional) – The size of the tiles. If not <cite>None</cite>, each scene is divided into square
-tiles of shape (tile_size, tile_size). The default is None.</p></li>
-<li><p><strong>pad</strong> (<cite>bool</cite>, optional) – Whether to center pad the input image. Set <code class="docutils literal notranslate"><span class="pre">pad</span></code> = True, if the
-images are not evenly divisible by the <code class="docutils literal notranslate"><span class="pre">tile_size</span></code>. The image data is
-padded with a constant padding value of zero. For each image, the
-corresponding ground truth image is padded with a “no data†label.
-The default is False.</p></li>
-<li><p><strong>gt_pattern</strong> (<cite>str</cite>, optional) – A regural expression to match the ground truth naming convention. All
-directories and subdirectories in <code class="docutils literal notranslate"><span class="pre">root_dir</span></code> are searched for files
-matching <code class="docutils literal notranslate"><span class="pre">gt_pattern</span></code>. The default is ‘(.*)gt\.tif’.</p></li>
-<li><p><strong>sort</strong> (<cite>bool</cite>, optional) – Whether to chronologically sort the samples. Useful for time series
-data. The default is False.</p></li>
-<li><p><strong>seed</strong> (<cite>int</cite>, optional) – The random seed. Used to split the dataset into training, validation
-and test set. Useful for reproducibility. The default is 0.</p></li>
-<li><p><strong>transforms</strong> (<cite>list</cite> [<cite>pysegcnn.core.split.Augment</cite>], optional) – List of <cite>pysegcnn.core.split.Augment</cite> instances. Each item in
-<code class="docutils literal notranslate"><span class="pre">transforms</span></code> generates a distinct transformed version of the dataset.
-The total dataset is composed of the original untransformed dataset
-together with each transformed version of it.
-If <code class="docutils literal notranslate"><span class="pre">transforms</span></code> = [], only the original dataset is used.
-The default is [].</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">build_samples</span></code>(scene)</p></td>
-<td><p>Stack the bands of a sample in a single array.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">compose_scenes</span></code>()</p></td>
-<td><p>Build the list of samples of the dataset.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_labels</span></code>()</p></td>
-<td><p>Class labels of the ProSnow datasets.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_sensor</span></code>()</p></td>
-<td><p>Sentinel 2 bands of the ProSnow datasets.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.dataset.ProSnowGarmisch.get_size" title="core.dataset.ProSnowGarmisch.get_size"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_size</span></code></a>()</p></td>
-<td><p>Image size of the ProSnow Garmisch dataset.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parse_scene_id</span></code>(scene_id)</p></td>
-<td><p>Parse ProSnow scene identifiers (Sentinel 2).</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">preprocess</span></code>(data, gt)</p></td>
-<td><p>Preprocess ProSnow dataset images.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">read_scene</span></code>(idx)</p></td>
-<td><p>Read the data of the sample with index <code class="docutils literal notranslate"><span class="pre">idx</span></code>.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to_tensor</span></code>(x, dtype)</p></td>
-<td><p>Convert <code class="docutils literal notranslate"><span class="pre">x</span></code> to <cite>torch.Tensor</cite>.</p></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="core.dataset.ProSnowGarmisch.get_size">
-<code class="sig-name descname">get_size</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#ProSnowGarmisch.get_size"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.ProSnowGarmisch.get_size" title="Permalink to this definition">¶</a></dt>
-<dd><p>Image size of the ProSnow Garmisch dataset.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>size</strong> – The image size (height, width).</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>tuple</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.dataset.ProSnowObergurgl">
-<em class="property">class </em><code class="sig-prename descclassname">core.dataset.</code><code class="sig-name descname">ProSnowObergurgl</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">root_dir</span></em>, <em class="sig-param"><span class="n">use_bands</span><span class="o">=</span><span class="default_value">[]</span></em>, <em class="sig-param"><span class="n">tile_size</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="n">pad</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">gt_pattern</span><span class="o">=</span><span class="default_value">'(.*)gt\\.tif'</span></em>, <em class="sig-param"><span class="n">sort</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">seed</span><span class="o">=</span><span class="default_value">0</span></em>, <em class="sig-param"><span class="n">transforms</span><span class="o">=</span><span class="default_value">[]</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#ProSnowObergurgl"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.ProSnowObergurgl" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.dataset.ProSnowDataset" title="core.dataset.ProSnowDataset"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.dataset.ProSnowDataset</span></code></a></p>
-<p>Class for the ProSnow Obergurgl dataset.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>root_dir</strong> (<cite>str</cite>) – The root directory, path to the dataset.</p></li>
-<li><p><strong>use_bands</strong> (<cite>list</cite> [<cite>str</cite>], optional) – A list of the spectral bands to use. The default is [].</p></li>
-<li><p><strong>tile_size</strong> (<cite>int</cite> or <cite>None</cite>, optional) – The size of the tiles. If not <cite>None</cite>, each scene is divided into square
-tiles of shape (tile_size, tile_size). The default is None.</p></li>
-<li><p><strong>pad</strong> (<cite>bool</cite>, optional) – Whether to center pad the input image. Set <code class="docutils literal notranslate"><span class="pre">pad</span></code> = True, if the
-images are not evenly divisible by the <code class="docutils literal notranslate"><span class="pre">tile_size</span></code>. The image data is
-padded with a constant padding value of zero. For each image, the
-corresponding ground truth image is padded with a “no data†label.
-The default is False.</p></li>
-<li><p><strong>gt_pattern</strong> (<cite>str</cite>, optional) – A regural expression to match the ground truth naming convention. All
-directories and subdirectories in <code class="docutils literal notranslate"><span class="pre">root_dir</span></code> are searched for files
-matching <code class="docutils literal notranslate"><span class="pre">gt_pattern</span></code>. The default is ‘(.*)gt\.tif’.</p></li>
-<li><p><strong>sort</strong> (<cite>bool</cite>, optional) – Whether to chronologically sort the samples. Useful for time series
-data. The default is False.</p></li>
-<li><p><strong>seed</strong> (<cite>int</cite>, optional) – The random seed. Used to split the dataset into training, validation
-and test set. Useful for reproducibility. The default is 0.</p></li>
-<li><p><strong>transforms</strong> (<cite>list</cite> [<cite>pysegcnn.core.split.Augment</cite>], optional) – List of <cite>pysegcnn.core.split.Augment</cite> instances. Each item in
-<code class="docutils literal notranslate"><span class="pre">transforms</span></code> generates a distinct transformed version of the dataset.
-The total dataset is composed of the original untransformed dataset
-together with each transformed version of it.
-If <code class="docutils literal notranslate"><span class="pre">transforms</span></code> = [], only the original dataset is used.
-The default is [].</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">build_samples</span></code>(scene)</p></td>
-<td><p>Stack the bands of a sample in a single array.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">compose_scenes</span></code>()</p></td>
-<td><p>Build the list of samples of the dataset.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_labels</span></code>()</p></td>
-<td><p>Class labels of the ProSnow datasets.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_sensor</span></code>()</p></td>
-<td><p>Sentinel 2 bands of the ProSnow datasets.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.dataset.ProSnowObergurgl.get_size" title="core.dataset.ProSnowObergurgl.get_size"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_size</span></code></a>()</p></td>
-<td><p>Image size of the ProSnow Obergurgl dataset.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parse_scene_id</span></code>(scene_id)</p></td>
-<td><p>Parse ProSnow scene identifiers (Sentinel 2).</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">preprocess</span></code>(data, gt)</p></td>
-<td><p>Preprocess ProSnow dataset images.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">read_scene</span></code>(idx)</p></td>
-<td><p>Read the data of the sample with index <code class="docutils literal notranslate"><span class="pre">idx</span></code>.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to_tensor</span></code>(x, dtype)</p></td>
-<td><p>Convert <code class="docutils literal notranslate"><span class="pre">x</span></code> to <cite>torch.Tensor</cite>.</p></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="core.dataset.ProSnowObergurgl.get_size">
-<code class="sig-name descname">get_size</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#ProSnowObergurgl.get_size"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.ProSnowObergurgl.get_size" title="Permalink to this definition">¶</a></dt>
-<dd><p>Image size of the ProSnow Obergurgl dataset.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>size</strong> – The image size (height, width).</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>tuple</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.dataset.SparcsDataset">
-<em class="property">class </em><code class="sig-prename descclassname">core.dataset.</code><code class="sig-name descname">SparcsDataset</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">root_dir</span></em>, <em class="sig-param"><span class="n">use_bands</span><span class="o">=</span><span class="default_value">[]</span></em>, <em class="sig-param"><span class="n">tile_size</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="n">pad</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">gt_pattern</span><span class="o">=</span><span class="default_value">'(.*)gt\\.tif'</span></em>, <em class="sig-param"><span class="n">sort</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">seed</span><span class="o">=</span><span class="default_value">0</span></em>, <em class="sig-param"><span class="n">transforms</span><span class="o">=</span><span class="default_value">[]</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#SparcsDataset"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.SparcsDataset" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.dataset.StandardEoDataset" title="core.dataset.StandardEoDataset"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.dataset.StandardEoDataset</span></code></a></p>
-<p>Class for the <a class="reference external" href="https://www.usgs.gov/land-resources/nli/landsat/spatial-procedures-automated-removal-cloud-and-shadow-sparcs-validation">Sparcs</a> dataset.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>root_dir</strong> (<cite>str</cite>) – The root directory, path to the dataset.</p></li>
-<li><p><strong>use_bands</strong> (<cite>list</cite> [<cite>str</cite>], optional) – A list of the spectral bands to use. The default is [].</p></li>
-<li><p><strong>tile_size</strong> (<cite>int</cite> or <cite>None</cite>, optional) – The size of the tiles. If not <cite>None</cite>, each scene is divided into square
-tiles of shape (tile_size, tile_size). The default is None.</p></li>
-<li><p><strong>pad</strong> (<cite>bool</cite>, optional) – Whether to center pad the input image. Set <code class="docutils literal notranslate"><span class="pre">pad</span></code> = True, if the
-images are not evenly divisible by the <code class="docutils literal notranslate"><span class="pre">tile_size</span></code>. The image data is
-padded with a constant padding value of zero. For each image, the
-corresponding ground truth image is padded with a “no data†label.
-The default is False.</p></li>
-<li><p><strong>gt_pattern</strong> (<cite>str</cite>, optional) – A regural expression to match the ground truth naming convention. All
-directories and subdirectories in <code class="docutils literal notranslate"><span class="pre">root_dir</span></code> are searched for files
-matching <code class="docutils literal notranslate"><span class="pre">gt_pattern</span></code>. The default is ‘(.*)gt\.tif’.</p></li>
-<li><p><strong>sort</strong> (<cite>bool</cite>, optional) – Whether to chronologically sort the samples. Useful for time series
-data. The default is False.</p></li>
-<li><p><strong>seed</strong> (<cite>int</cite>, optional) – The random seed. Used to split the dataset into training, validation
-and test set. Useful for reproducibility. The default is 0.</p></li>
-<li><p><strong>transforms</strong> (<cite>list</cite> [<cite>pysegcnn.core.split.Augment</cite>], optional) – List of <cite>pysegcnn.core.split.Augment</cite> instances. Each item in
-<code class="docutils literal notranslate"><span class="pre">transforms</span></code> generates a distinct transformed version of the dataset.
-The total dataset is composed of the original untransformed dataset
-together with each transformed version of it.
-If <code class="docutils literal notranslate"><span class="pre">transforms</span></code> = [], only the original dataset is used.
-The default is [].</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">build_samples</span></code>(scene)</p></td>
-<td><p>Stack the bands of a sample in a single array.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">compose_scenes</span></code>()</p></td>
-<td><p>Build the list of samples of the dataset.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.dataset.SparcsDataset.get_labels" title="core.dataset.SparcsDataset.get_labels"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_labels</span></code></a>()</p></td>
-<td><p>Class labels of the Sparcs dataset.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.dataset.SparcsDataset.get_sensor" title="core.dataset.SparcsDataset.get_sensor"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_sensor</span></code></a>()</p></td>
-<td><p>Landsat 8 bands of the Sparcs dataset.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.dataset.SparcsDataset.get_size" title="core.dataset.SparcsDataset.get_size"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_size</span></code></a>()</p></td>
-<td><p>Image size of the Sparcs dataset.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.dataset.SparcsDataset.parse_scene_id" title="core.dataset.SparcsDataset.parse_scene_id"><code class="xref py py-obj docutils literal notranslate"><span class="pre">parse_scene_id</span></code></a>(scene_id)</p></td>
-<td><p>Parse Sparcs scene identifiers (Landsat 8).</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.dataset.SparcsDataset.preprocess" title="core.dataset.SparcsDataset.preprocess"><code class="xref py py-obj docutils literal notranslate"><span class="pre">preprocess</span></code></a>(data, gt)</p></td>
-<td><p>Preprocess Sparcs dataset images.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">read_scene</span></code>(idx)</p></td>
-<td><p>Read the data of the sample with index <code class="docutils literal notranslate"><span class="pre">idx</span></code>.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to_tensor</span></code>(x, dtype)</p></td>
-<td><p>Convert <code class="docutils literal notranslate"><span class="pre">x</span></code> to <cite>torch.Tensor</cite>.</p></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="core.dataset.SparcsDataset.get_labels">
-<code class="sig-name descname">get_labels</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#SparcsDataset.get_labels"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.SparcsDataset.get_labels" title="Permalink to this definition">¶</a></dt>
-<dd><p>Class labels of the Sparcs dataset.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>labels</strong> – The class labels.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>enum.Enum</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.dataset.SparcsDataset.get_sensor">
-<code class="sig-name descname">get_sensor</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#SparcsDataset.get_sensor"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.SparcsDataset.get_sensor" title="Permalink to this definition">¶</a></dt>
-<dd><p>Landsat 8 bands of the Sparcs dataset.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>sensor</strong> – An enumeration of the bands of the sensor.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>enum.Enum</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.dataset.SparcsDataset.get_size">
-<code class="sig-name descname">get_size</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#SparcsDataset.get_size"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.SparcsDataset.get_size" title="Permalink to this definition">¶</a></dt>
-<dd><p>Image size of the Sparcs dataset.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>size</strong> – The image size (height, width).</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>tuple</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.dataset.SparcsDataset.parse_scene_id">
-<code class="sig-name descname">parse_scene_id</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">scene_id</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#SparcsDataset.parse_scene_id"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.SparcsDataset.parse_scene_id" title="Permalink to this definition">¶</a></dt>
-<dd><p>Parse Sparcs scene identifiers (Landsat 8).</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>scene_id</strong> (<cite>str</cite>) – A scene identifier.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>scene</strong> – A dictionary containing scene metadata. If <cite>None</cite>, <code class="docutils literal notranslate"><span class="pre">scene_id</span></code> is
-not a valid Landsat scene identifier.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>dict</cite> or <cite>None</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.dataset.SparcsDataset.preprocess">
-<code class="sig-name descname">preprocess</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">data</span></em>, <em class="sig-param"><span class="n">gt</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#SparcsDataset.preprocess"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.SparcsDataset.preprocess" title="Permalink to this definition">¶</a></dt>
-<dd><p>Preprocess Sparcs dataset images.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>data</strong> (<cite>numpy.ndarray</cite>) – The sample input data.</p></li>
-<li><p><strong>gt</strong> (<cite>numpy.ndarray</cite>) – The sample ground truth.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><ul class="simple">
-<li><p><strong>data</strong> (<cite>numpy.ndarray</cite>) – The preprocessed input data.</p></li>
-<li><p><strong>gt</strong> (<cite>numpy.ndarray</cite>) – The preprocessed ground truth data.</p></li>
-</ul>
-</p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.dataset.StandardEoDataset">
-<em class="property">class </em><code class="sig-prename descclassname">core.dataset.</code><code class="sig-name descname">StandardEoDataset</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">root_dir</span></em>, <em class="sig-param"><span class="n">use_bands</span><span class="o">=</span><span class="default_value">[]</span></em>, <em class="sig-param"><span class="n">tile_size</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="n">pad</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">gt_pattern</span><span class="o">=</span><span class="default_value">'(.*)gt\\.tif'</span></em>, <em class="sig-param"><span class="n">sort</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">seed</span><span class="o">=</span><span class="default_value">0</span></em>, <em class="sig-param"><span class="n">transforms</span><span class="o">=</span><span class="default_value">[]</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#StandardEoDataset"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.StandardEoDataset" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.dataset.ImageDataset" title="core.dataset.ImageDataset"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.dataset.ImageDataset</span></code></a></p>
-<p>Base class for standard Earth Observation style datasets.</p>
-<p><cite>pysegcnn.core.dataset.StandardEoDataset</cite> implements the
-<cite>~pysegcnn.core.dataset.StandardEoDataset.compose_scenes</cite> method for
-datasets with the following directory structure:</p>
-<dl>
-<dt>root_dir/</dt><dd><dl class="simple">
-<dt>scene_id_1/</dt><dd><p>scene_id_1_B1.tif
-scene_id_1_B2.tif
-.
-.
-.
-scene_id_1_BN.tif</p>
-</dd>
-<dt>scene_id_2/</dt><dd><p>scene_id_2_B1.tif
-scene_id_2_B2.tif
-.
-.
-.
-scene_id_2_BN.tif</p>
-</dd>
-</dl>
-<p>.
-scene_id_N/</p>
-<blockquote>
-<div><p>.</p>
-</div></blockquote>
-</dd>
-</dl>
-<p>If your dataset shares this directory structure, you can directly inherit
-<cite>pysegcnn.core.dataset.StandardEoDataset</cite> and implement the remaining
-methods.</p>
-<p>See <cite>pysegcnn.core.dataset.SparcsDataset</cite> for an example.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>root_dir</strong> (<cite>str</cite>) – The root directory, path to the dataset.</p></li>
-<li><p><strong>use_bands</strong> (<cite>list</cite> [<cite>str</cite>], optional) – A list of the spectral bands to use. The default is [].</p></li>
-<li><p><strong>tile_size</strong> (<cite>int</cite> or <cite>None</cite>, optional) – The size of the tiles. If not <cite>None</cite>, each scene is divided into square
-tiles of shape (tile_size, tile_size). The default is None.</p></li>
-<li><p><strong>pad</strong> (<cite>bool</cite>, optional) – Whether to center pad the input image. Set <code class="docutils literal notranslate"><span class="pre">pad</span></code> = True, if the
-images are not evenly divisible by the <code class="docutils literal notranslate"><span class="pre">tile_size</span></code>. The image data is
-padded with a constant padding value of zero. For each image, the
-corresponding ground truth image is padded with a “no data†label.
-The default is False.</p></li>
-<li><p><strong>gt_pattern</strong> (<cite>str</cite>, optional) – A regural expression to match the ground truth naming convention. All
-directories and subdirectories in <code class="docutils literal notranslate"><span class="pre">root_dir</span></code> are searched for files
-matching <code class="docutils literal notranslate"><span class="pre">gt_pattern</span></code>. The default is ‘(.*)gt\.tif’.</p></li>
-<li><p><strong>sort</strong> (<cite>bool</cite>, optional) – Whether to chronologically sort the samples. Useful for time series
-data. The default is False.</p></li>
-<li><p><strong>seed</strong> (<cite>int</cite>, optional) – The random seed. Used to split the dataset into training, validation
-and test set. Useful for reproducibility. The default is 0.</p></li>
-<li><p><strong>transforms</strong> (<cite>list</cite> [<cite>pysegcnn.core.split.Augment</cite>], optional) – List of <cite>pysegcnn.core.split.Augment</cite> instances. Each item in
-<code class="docutils literal notranslate"><span class="pre">transforms</span></code> generates a distinct transformed version of the dataset.
-The total dataset is composed of the original untransformed dataset
-together with each transformed version of it.
-If <code class="docutils literal notranslate"><span class="pre">transforms</span></code> = [], only the original dataset is used.
-The default is [].</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">build_samples</span></code>(scene)</p></td>
-<td><p>Stack the bands of a sample in a single array.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.dataset.StandardEoDataset.compose_scenes" title="core.dataset.StandardEoDataset.compose_scenes"><code class="xref py py-obj docutils literal notranslate"><span class="pre">compose_scenes</span></code></a>()</p></td>
-<td><p>Build the list of samples of the dataset.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_labels</span></code>()</p></td>
-<td><p>Return an enumeration of the class labels of the dataset.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_sensor</span></code>()</p></td>
-<td><p>Return an enumeration of the bands of the sensor of the dataset.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_size</span></code>()</p></td>
-<td><p>Return the size of the images in the dataset.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parse_scene_id</span></code>(scene_id)</p></td>
-<td><p>Parse the scene identifier.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">preprocess</span></code>(data, gt)</p></td>
-<td><p>Preprocess a sample before feeding it to a model.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">read_scene</span></code>(idx)</p></td>
-<td><p>Read the data of the sample with index <code class="docutils literal notranslate"><span class="pre">idx</span></code>.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to_tensor</span></code>(x, dtype)</p></td>
-<td><p>Convert <code class="docutils literal notranslate"><span class="pre">x</span></code> to <cite>torch.Tensor</cite>.</p></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="core.dataset.StandardEoDataset.compose_scenes">
-<code class="sig-name descname">compose_scenes</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#StandardEoDataset.compose_scenes"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.StandardEoDataset.compose_scenes" title="Permalink to this definition">¶</a></dt>
-<dd><p>Build the list of samples of the dataset.</p>
-<p>Each sample is represented by a dictionary.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><p><strong>scenes</strong> –</p>
-<dl class="simple">
-<dt>Each item in <code class="docutils literal notranslate"><span class="pre">scenes</span></code> is a <cite>dict</cite> with keys:</dt><dd><dl class="simple">
-<dt><code class="docutils literal notranslate"><span class="pre">'band_name_1'</span></code></dt><dd><p>Path to the file of band_1.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'band_name_2'</span></code></dt><dd><p>Path to the file of band_2.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'band_name_n'</span></code></dt><dd><p>Path to the file of band_n.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'gt'</span></code></dt><dd><p>Path to the ground truth file.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'date'</span></code></dt><dd><p>The date of the sample.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'tile'</span></code></dt><dd><p>The tile id of the sample.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'transform'</span></code></dt><dd><p>The transformation to apply.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'id'</span></code></dt><dd><p>The scene identifier.</p>
-</dd>
-</dl>
-</dd>
-</dl>
-</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>list</cite> [<cite>dict</cite>]</p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.dataset.SupportedDatasets">
-<em class="property">class </em><code class="sig-prename descclassname">core.dataset.</code><code class="sig-name descname">SupportedDatasets</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">value</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/dataset.html#SupportedDatasets"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.dataset.SupportedDatasets" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">enum.Enum</span></code></p>
-<p>Names and corresponding classes of the implemented datasets.</p>
-<dl class="py attribute">
-<dt id="core.dataset.SupportedDatasets.Cloud95">
-<code class="sig-name descname">Cloud95</code><em class="property"> = <class 'core.dataset.Cloud95Dataset'></em><a class="headerlink" href="#core.dataset.SupportedDatasets.Cloud95" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.dataset.SupportedDatasets.Garmisch">
-<code class="sig-name descname">Garmisch</code><em class="property"> = <class 'core.dataset.ProSnowGarmisch'></em><a class="headerlink" href="#core.dataset.SupportedDatasets.Garmisch" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.dataset.SupportedDatasets.Obergurgl">
-<code class="sig-name descname">Obergurgl</code><em class="property"> = <class 'core.dataset.ProSnowObergurgl'></em><a class="headerlink" href="#core.dataset.SupportedDatasets.Obergurgl" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.dataset.SupportedDatasets.Sparcs">
-<code class="sig-name descname">Sparcs</code><em class="property"> = <class 'core.dataset.SparcsDataset'></em><a class="headerlink" href="#core.dataset.SupportedDatasets.Sparcs" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-</dd></dl>
-
-</div>
-</div>
-<div class="section" id="module-core.graphics">
-<span id="core-graphics-module"></span><h2>core.graphics module<a class="headerlink" href="#module-core.graphics" title="Permalink to this headline">¶</a></h2>
-<p>Functions to plot multispectral image data and model output.</p>
-<div class="section" id="id4">
-<h3>License<a class="headerlink" href="#id4" title="Permalink to this headline">¶</a></h3>
-<blockquote>
-<div><p>Copyright (c) 2020 Daniel Frisinghelli</p>
-<p>This source code is licensed under the GNU General Public License v3.</p>
-<p>See the LICENSE file in the repository’s root directory.</p>
-</div></blockquote>
-<dl class="py function">
-<dt id="core.graphics.contrast_stretching">
-<code class="sig-prename descclassname">core.graphics.</code><code class="sig-name descname">contrast_stretching</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">image</span></em>, <em class="sig-param"><span class="n">alpha</span><span class="o">=</span><span class="default_value">5</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/graphics.html#contrast_stretching"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.graphics.contrast_stretching" title="Permalink to this definition">¶</a></dt>
-<dd><p>Apply percentile stretching to an image to increase constrast.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>image</strong> (<cite>numpy.ndarray</cite>) – the input image.</p></li>
-<li><p><strong>alpha</strong> (<cite>int</cite>, optional) – The level of the percentiles. The default is 5.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>norm</strong> – the stretched image.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>numpy.ndarray</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt id="core.graphics.plot_confusion_matrix">
-<code class="sig-prename descclassname">core.graphics.</code><code class="sig-name descname">plot_confusion_matrix</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">cm</span></em>, <em class="sig-param"><span class="n">labels</span></em>, <em class="sig-param"><span class="n">normalize</span><span class="o">=</span><span class="default_value">True</span></em>, <em class="sig-param"><span class="n">figsize</span><span class="o">=</span><span class="default_value">10, 10</span></em>, <em class="sig-param"><span class="n">cmap</span><span class="o">=</span><span class="default_value">'Blues'</span></em>, <em class="sig-param"><span class="n">state</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="n">outpath</span><span class="o">=</span><span class="default_value">'c:\\eurac\\2020\\git\\pysegcnn\\pysegcnn\\main\\_graphics/'</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/graphics.html#plot_confusion_matrix"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.graphics.plot_confusion_matrix" title="Permalink to this definition">¶</a></dt>
-<dd><p>Plot the confusion matrix <code class="docutils literal notranslate"><span class="pre">cm</span></code>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>cm</strong> (<cite>numpy.ndarray</cite>) – The confusion matrix.</p></li>
-<li><p><strong>labels</strong> (<cite>dict</cite> [<cite>int</cite>, <cite>dict</cite>]) – <p>The label dictionary. The keys are the values of the class labels
-in the ground truth <code class="docutils literal notranslate"><span class="pre">y</span></code>. Each nested <cite>dict</cite> should have keys:</p>
-<blockquote>
-<div><dl class="simple">
-<dt><code class="docutils literal notranslate"><span class="pre">'color'</span></code></dt><dd><p>A named color (<cite>str</cite>).</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'label'</span></code></dt><dd><p>The name of the class label (<cite>str</cite>).</p>
-</dd>
-</dl>
-</div></blockquote>
-</p></li>
-<li><p><strong>normalize</strong> (<cite>bool</cite>, optional) – Whether to normalize the confusion matrix. The default is True.</p></li>
-<li><p><strong>figsize</strong> (<cite>tuple</cite>, optional) – The figure size in centimeters. The default is (10, 10).</p></li>
-<li><p><strong>cmap</strong> (<cite>str</cite>, optional) – A colormap in <cite>matplotlib.pyplot.colormaps()</cite>. The default is ‘Blues’.</p></li>
-<li><p><strong>state</strong> (<cite>str</cite> or <cite>None</cite>, optional) – Filename to save the plot to. <code class="docutils literal notranslate"><span class="pre">state</span></code> should be an existing model
-state file ending with ‘.pt’. The default is None, i.e. plot is not
-saved to disk.</p></li>
-<li><p><strong>outpath</strong> (<cite>str</cite> or <cite>pathlib.Path</cite>, optional) – Output path. The default is ‘pysegcnn/main/_graphics/’.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><ul class="simple">
-<li><p><strong>fig</strong> (<cite>matplotlib.figure.Figure</cite>) – The figure handle.</p></li>
-<li><p><strong>ax</strong> (<cite>matplotlib.axes._subplots.AxesSubplot</cite>) – The axes handle.</p></li>
-</ul>
-</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt id="core.graphics.plot_loss">
-<code class="sig-prename descclassname">core.graphics.</code><code class="sig-name descname">plot_loss</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">state_file</span></em>, <em class="sig-param"><span class="n">figsize</span><span class="o">=</span><span class="default_value">10, 10</span></em>, <em class="sig-param"><span class="n">step</span><span class="o">=</span><span class="default_value">5</span></em>, <em class="sig-param"><span class="n">colors</span><span class="o">=</span><span class="default_value">['lightgreen', 'green', 'skyblue', 'steelblue']</span></em>, <em class="sig-param"><span class="n">outpath</span><span class="o">=</span><span class="default_value">'c:\\eurac\\2020\\git\\pysegcnn\\pysegcnn\\main\\_graphics/'</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/graphics.html#plot_loss"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.graphics.plot_loss" title="Permalink to this definition">¶</a></dt>
-<dd><p>Plot the observed loss and accuracy of a model run.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>state_file</strong> (<cite>str</cite> or <cite>pathlib.Path</cite>) – The model state file. Model state files are stored in
-pysegcnn/main/_models.</p></li>
-<li><p><strong>figsize</strong> (<cite>tuple</cite>, optional) – The figure size in centimeters. The default is (10, 10).</p></li>
-<li><p><strong>step</strong> (<cite>int</cite>, optional) – The step of epochs for the x-axis labels. The default is 5, i.e. label
-each fifth epoch.</p></li>
-<li><p><strong>colors</strong> (<cite>list</cite> [<cite>str</cite>], optional) – A list of four named colors supported by <cite>matplotlib</cite>.
-The default is [‘lightgreen’, ‘green’, ‘skyblue’, ‘steelblue’].</p></li>
-<li><p><strong>outpath</strong> (<cite>str</cite> or <cite>pathlib.Path</cite>, optional) – Output path. The default is ‘pysegcnn/main/_graphics/’.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>fig</strong> – The figure handle.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>matplotlib.figure.Figure</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt id="core.graphics.plot_sample">
-<code class="sig-prename descclassname">core.graphics.</code><code class="sig-name descname">plot_sample</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">x</span></em>, <em class="sig-param"><span class="n">use_bands</span></em>, <em class="sig-param"><span class="n">labels</span></em>, <em class="sig-param"><span class="n">y</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="n">y_pred</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="n">figsize</span><span class="o">=</span><span class="default_value">10, 10</span></em>, <em class="sig-param"><span class="n">bands</span><span class="o">=</span><span class="default_value">['nir', 'red', 'green']</span></em>, <em class="sig-param"><span class="n">state</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="n">outpath</span><span class="o">=</span><span class="default_value">'c:\\eurac\\2020\\git\\pysegcnn\\pysegcnn\\main\\_samples/'</span></em>, <em class="sig-param"><span class="n">alpha</span><span class="o">=</span><span class="default_value">0</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/graphics.html#plot_sample"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.graphics.plot_sample" title="Permalink to this definition">¶</a></dt>
-<dd><p>Plot false color composite (FCC), ground truth and model prediction.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>x</strong> (<cite>numpy.ndarray</cite> or <cite>torch.Tensor</cite>, (b, h, w)) – Array containing the raw data of the tile, shape=(bands, height, width)</p></li>
-<li><p><strong>use_bands</strong> (<cite>list</cite> of <cite>str</cite>) – List describing the order of the bands in <code class="docutils literal notranslate"><span class="pre">x</span></code>.</p></li>
-<li><p><strong>labels</strong> (<cite>dict</cite> [<cite>int</cite>, <cite>dict</cite>]) – <p>The label dictionary. The keys are the values of the class labels
-in the ground truth <code class="docutils literal notranslate"><span class="pre">y</span></code>. Each nested <cite>dict</cite> should have keys:</p>
-<blockquote>
-<div><dl class="simple">
-<dt><code class="docutils literal notranslate"><span class="pre">'color'</span></code></dt><dd><p>A named color (<cite>str</cite>).</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'label'</span></code></dt><dd><p>The name of the class label (<cite>str</cite>).</p>
-</dd>
-</dl>
-</div></blockquote>
-</p></li>
-<li><p><strong>y</strong> (<cite>numpy.ndarray</cite> or <cite>torch.Tensor</cite> or <cite>None</cite>, optional) – Array containing the ground truth of tile <code class="docutils literal notranslate"><span class="pre">x</span></code>, shape=(height, width).
-The default is None.</p></li>
-<li><p><strong>y_pred</strong> (<cite>numpy.ndarray</cite> or <cite>torch.Tensor</cite> or <cite>None</cite>, optional) – Array containing the prediction for tile <code class="docutils literal notranslate"><span class="pre">x</span></code>, shape=(height, width).
-The default is None.</p></li>
-<li><p><strong>figsize</strong> (<cite>tuple</cite>, optional) – The figure size in centimeters. The default is (10, 10).</p></li>
-<li><p><strong>bands</strong> (<cite>list</cite> [<cite>str</cite>], optional) – The bands to build the FCC. The default is [‘nir’, ‘red’, ‘green’].</p></li>
-<li><p><strong>state</strong> (<cite>str</cite> or <cite>None</cite>, optional) – Filename to save the plot to. <code class="docutils literal notranslate"><span class="pre">state</span></code> should be an existing model
-state file ending with ‘.pt’. The default is None, i.e. plot is not
-saved to disk.</p></li>
-<li><p><strong>outpath</strong> (<cite>str</cite> or <cite>pathlib.Path</cite>, optional) – Output path. The default is ‘pysegcnn/main/_samples’.</p></li>
-<li><p><strong>alpha</strong> (<cite>int</cite>, optional) – The level of the percentiles to increase constrast in the FCC.
-The default is 0, i.e. no stretching.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><ul class="simple">
-<li><p><strong>fig</strong> (<cite>matplotlib.figure.Figure</cite>) – The figure handle.</p></li>
-<li><p><strong>ax</strong> (<cite>numpy.ndarray</cite> [<cite>matplotlib.axes._subplots.AxesSubplot</cite>]) – An array of the axes handles.</p></li>
-</ul>
-</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt id="core.graphics.running_mean">
-<code class="sig-prename descclassname">core.graphics.</code><code class="sig-name descname">running_mean</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">x</span></em>, <em class="sig-param"><span class="n">w</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/graphics.html#running_mean"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.graphics.running_mean" title="Permalink to this definition">¶</a></dt>
-<dd><p>Compute a running mean of the input sequence.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>x</strong> (<em>array_like</em>) – The sequence to compute a running mean on.</p></li>
-<li><p><strong>w</strong> (<cite>int</cite>) – The window length of the running mean.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>rm</strong> – The running mean of the sequence <code class="docutils literal notranslate"><span class="pre">x</span></code>.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>numpy.ndarray</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-</div>
-</div>
-<div class="section" id="module-core.layers">
-<span id="core-layers-module"></span><h2>core.layers module<a class="headerlink" href="#module-core.layers" title="Permalink to this headline">¶</a></h2>
-<p>Layers of a convolutional encoder-decoder network.</p>
-<div class="section" id="id5">
-<h3>License<a class="headerlink" href="#id5" title="Permalink to this headline">¶</a></h3>
-<blockquote>
-<div><p>Copyright (c) 2020 Daniel Frisinghelli</p>
-<p>This source code is licensed under the GNU General Public License v3.</p>
-<p>See the LICENSE file in the repository’s root directory.</p>
-</div></blockquote>
-<dl class="py class">
-<dt id="core.layers.Block">
-<em class="property">class </em><code class="sig-prename descclassname">core.layers.</code><code class="sig-name descname">Block</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">out_channels</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#Block"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.Block" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">torch.nn.modules.module.Module</span></code></p>
-<p>Basic convolutional block.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>in_channels</strong> (<cite>int</cite>) – Number of input channels.</p></li>
-<li><p><strong>out_channels</strong> (<cite>int</cite>) – Number of output channels.</p></li>
-<li><p><strong>**kwargs</strong> (‘dict’ [<cite>str</cite>]) – Additional arguments passed to <cite>pysegcnn.core.layers.Conv2dSame</cite>.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Raises</dt>
-<dd class="field-even"><p><strong>TypeError</strong> – Raised if <cite>~pysegcnn.core.layers.Block.layers</cite> method does not return
-an instance of <cite>torch.nn.Sequential</cite>.</p>
-</dd>
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p></p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
-<td><p>Adds a child module to the current module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">apply</span></code>(fn)</p></td>
-<td><p>Applies <code class="docutils literal notranslate"><span class="pre">fn</span></code> recursively to every submodule (as returned by <code class="docutils literal notranslate"><span class="pre">.children()</span></code>) as well as self.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">bfloat16</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">bfloat16</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">buffers</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cpu</span></code>()</p></td>
-<td><p>Moves all model parameters and buffers to the CPU.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cuda</span></code>([device])</p></td>
-<td><p>Moves all model parameters and buffers to the GPU.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">double</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">double</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">eval</span></code>()</p></td>
-<td><p>Sets the module in evaluation mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">extra_repr</span></code>()</p></td>
-<td><p>Set the extra representation of the module</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">float</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to float datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.layers.Block.forward" title="core.layers.Block.forward"><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward</span></code></a>()</p></td>
-<td><p>Forward pass of the block.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">half</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">half</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.layers.Block.layers" title="core.layers.Block.layers"><code class="xref py py-obj docutils literal notranslate"><span class="pre">layers</span></code></a>()</p></td>
-<td><p>Define the layers of the block.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_state_dict</span></code>(state_dict[, strict])</p></td>
-<td><p>Copies parameters and buffers from <code class="xref py py-attr docutils literal notranslate"><span class="pre">state_dict</span></code> into this module and its descendants.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">modules</span></code>()</p></td>
-<td><p>Returns an iterator over all modules in the network.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_buffers</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_modules</span></code>([memo, prefix])</p></td>
-<td><p>Returns an iterator over all modules in the network, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_parameters</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parameters</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module parameters.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_backward_hook</span></code>(hook)</p></td>
-<td><p>Registers a backward hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_buffer</span></code>(name, tensor[, persistent])</p></td>
-<td><p>Adds a buffer to the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_pre_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward pre-hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_parameter</span></code>(name, param)</p></td>
-<td><p>Adds a parameter to the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">requires_grad_</span></code>([requires_grad])</p></td>
-<td><p>Change if autograd should record operations on parameters in this module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">state_dict</span></code>([destination, prefix, keep_vars])</p></td>
-<td><p>Returns a dictionary containing a whole state of the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to</span></code>(*args, **kwargs)</p></td>
-<td><p>Moves and/or casts the parameters and buffers.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">train</span></code>([mode])</p></td>
-<td><p>Sets the module in training mode.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">type</span></code>(dst_type)</p></td>
-<td><p>Casts all parameters and buffers to <code class="xref py py-attr docutils literal notranslate"><span class="pre">dst_type</span></code>.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">zero_grad</span></code>()</p></td>
-<td><p>Sets gradients of all model parameters to zero.</p></td>
-</tr>
-</tbody>
-</table>
-<table class="docutils align-default">
-<colgroup>
-<col style="width: 62%" />
-<col style="width: 38%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><strong>__call__</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-even"><td><p><strong>share_memory</strong></p></td>
-<td></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="core.layers.Block.forward">
-<code class="sig-name descname">forward</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#Block.forward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.Block.forward" title="Permalink to this definition">¶</a></dt>
-<dd><p>Forward pass of the block.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Raises</dt>
-<dd class="field-odd"><p><strong>NotImplementedError</strong> – Raised if <cite>pysegcnn.core.layers.Block</cite> is not inherited.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.layers.Block.layers">
-<code class="sig-name descname">layers</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#Block.layers"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.Block.layers" title="Permalink to this definition">¶</a></dt>
-<dd><p>Define the layers of the block.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Raises</dt>
-<dd class="field-odd"><p><strong>NotImplementedError</strong> – Raised if <cite>pysegcnn.core.layers.Block</cite> is not inherited.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>layers</strong> – Return an instance of <cite>torch.nn.Sequential</cite> containing a sequence
-of layer (<cite>torch.nn.Module</cite> ) instances.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>torch.nn.Sequential</cite> [<cite>torch.nn.Module</cite>]</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt id="core.layers.Block.training">
-<code class="sig-name descname">training</code><em class="property">: bool</em><a class="headerlink" href="#core.layers.Block.training" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.layers.Conv2dSame">
-<em class="property">class </em><code class="sig-prename descclassname">core.layers.</code><code class="sig-name descname">Conv2dSame</code><span class="sig-paren">(</span><em class="sig-param"><span class="o">*</span><span class="n">args</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#Conv2dSame"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.Conv2dSame" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">torch.nn.modules.conv.Conv2d</span></code></p>
-<p>A convolution preserving the shape of its input.</p>
-<p>Given the kernel size, the dilation and a stride of 1, the padding is
-calculated such that the output of the convolution has the same spatial
-dimensions as the input.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>*args</strong> (<cite>list</cite> [<cite>str</cite>]) – <dl class="simple">
-<dt>positional arguments passed to <cite>torch.nn.Conv2d</cite>:</dt><dd><dl class="simple">
-<dt><code class="docutils literal notranslate"><span class="pre">'in_channels'</span></code>: <cite>int</cite></dt><dd><p>Number of input channels.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'out_channels'</span></code>: <cite>int</cite></dt><dd><p>Number of output channels.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'kernel_size'</span></code>: <cite>int</cite> or <cite>tuple</cite> [<cite>int</cite>]</dt><dd><p>Size of the convolving kernel.</p>
-</dd>
-</dl>
-</dd>
-</dl>
-</p></li>
-<li><p><strong>**kwargs</strong> (‘dict’ [<cite>str</cite>]) – Additional keyword arguments passed to <a href="#id23"><span class="problematic" id="id24">`torch.nn.Conv2d`_</span></a>.</p></li>
-<li><p><strong>_torch.nn.Conv2d</strong> (<em>.</em>) – <a class="reference external" href="https://pytorch.org/docs/stable/generated/torch.nn.Conv2d.html#torch.nn.Conv2d">https://pytorch.org/docs/stable/generated/torch.nn.Conv2d.html#torch.nn.Conv2d</a></p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
-<td><p>Adds a child module to the current module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">apply</span></code>(fn)</p></td>
-<td><p>Applies <code class="docutils literal notranslate"><span class="pre">fn</span></code> recursively to every submodule (as returned by <code class="docutils literal notranslate"><span class="pre">.children()</span></code>) as well as self.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">bfloat16</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">bfloat16</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">buffers</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cpu</span></code>()</p></td>
-<td><p>Moves all model parameters and buffers to the CPU.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cuda</span></code>([device])</p></td>
-<td><p>Moves all model parameters and buffers to the GPU.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">double</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">double</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">eval</span></code>()</p></td>
-<td><p>Sets the module in evaluation mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">extra_repr</span></code>()</p></td>
-<td><p>Set the extra representation of the module</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">float</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to float datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">half</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">half</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_state_dict</span></code>(state_dict[, strict])</p></td>
-<td><p>Copies parameters and buffers from <code class="xref py py-attr docutils literal notranslate"><span class="pre">state_dict</span></code> into this module and its descendants.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">modules</span></code>()</p></td>
-<td><p>Returns an iterator over all modules in the network.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_buffers</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_modules</span></code>([memo, prefix])</p></td>
-<td><p>Returns an iterator over all modules in the network, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_parameters</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parameters</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module parameters.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_backward_hook</span></code>(hook)</p></td>
-<td><p>Registers a backward hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_buffer</span></code>(name, tensor[, persistent])</p></td>
-<td><p>Adds a buffer to the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_pre_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward pre-hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_parameter</span></code>(name, param)</p></td>
-<td><p>Adds a parameter to the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">requires_grad_</span></code>([requires_grad])</p></td>
-<td><p>Change if autograd should record operations on parameters in this module.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.layers.Conv2dSame.same_padding" title="core.layers.Conv2dSame.same_padding"><code class="xref py py-obj docutils literal notranslate"><span class="pre">same_padding</span></code></a>(d, k)</p></td>
-<td><p>Calculate the amount of padding.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">state_dict</span></code>([destination, prefix, keep_vars])</p></td>
-<td><p>Returns a dictionary containing a whole state of the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to</span></code>(*args, **kwargs)</p></td>
-<td><p>Moves and/or casts the parameters and buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">train</span></code>([mode])</p></td>
-<td><p>Sets the module in training mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">type</span></code>(dst_type)</p></td>
-<td><p>Casts all parameters and buffers to <code class="xref py py-attr docutils literal notranslate"><span class="pre">dst_type</span></code>.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">zero_grad</span></code>()</p></td>
-<td><p>Sets gradients of all model parameters to zero.</p></td>
-</tr>
-</tbody>
-</table>
-<table class="docutils align-default">
-<colgroup>
-<col style="width: 67%" />
-<col style="width: 33%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><strong>__call__</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-even"><td><p><strong>forward</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-odd"><td><p><strong>reset_parameters</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-even"><td><p><strong>share_memory</strong></p></td>
-<td></td>
-</tr>
-</tbody>
-</table>
-<dl class="py attribute">
-<dt id="core.layers.Conv2dSame.bias">
-<code class="sig-name descname">bias</code><em class="property">: Optional<span class="p">[</span>torch.Tensor<span class="p">]</span></em><a class="headerlink" href="#core.layers.Conv2dSame.bias" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.layers.Conv2dSame.dilation">
-<code class="sig-name descname">dilation</code><em class="property">: Tuple<span class="p">[</span>int<span class="p">, </span><span class="p">…</span><span class="p">]</span></em><a class="headerlink" href="#core.layers.Conv2dSame.dilation" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.layers.Conv2dSame.groups">
-<code class="sig-name descname">groups</code><em class="property">: int</em><a class="headerlink" href="#core.layers.Conv2dSame.groups" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.layers.Conv2dSame.kernel_size">
-<code class="sig-name descname">kernel_size</code><em class="property">: Tuple<span class="p">[</span>int<span class="p">, </span><span class="p">…</span><span class="p">]</span></em><a class="headerlink" href="#core.layers.Conv2dSame.kernel_size" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.layers.Conv2dSame.out_channels">
-<code class="sig-name descname">out_channels</code><em class="property">: int</em><a class="headerlink" href="#core.layers.Conv2dSame.out_channels" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.layers.Conv2dSame.output_padding">
-<code class="sig-name descname">output_padding</code><em class="property">: Tuple<span class="p">[</span>int<span class="p">, </span><span class="p">…</span><span class="p">]</span></em><a class="headerlink" href="#core.layers.Conv2dSame.output_padding" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.layers.Conv2dSame.padding">
-<code class="sig-name descname">padding</code><em class="property">: Tuple<span class="p">[</span>int<span class="p">, </span><span class="p">…</span><span class="p">]</span></em><a class="headerlink" href="#core.layers.Conv2dSame.padding" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.layers.Conv2dSame.padding_mode">
-<code class="sig-name descname">padding_mode</code><em class="property">: str</em><a class="headerlink" href="#core.layers.Conv2dSame.padding_mode" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py method">
-<dt id="core.layers.Conv2dSame.same_padding">
-<em class="property">static </em><code class="sig-name descname">same_padding</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">d</span></em>, <em class="sig-param"><span class="n">k</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#Conv2dSame.same_padding"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.Conv2dSame.same_padding" title="Permalink to this definition">¶</a></dt>
-<dd><p>Calculate the amount of padding.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>d</strong> (<cite>int</cite>) – The dilation of the convolution.</p></li>
-<li><p><strong>k</strong> (<cite>int</cite>) – The kernel size.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>p</strong> – the amount of padding.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>int</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt id="core.layers.Conv2dSame.stride">
-<code class="sig-name descname">stride</code><em class="property">: Tuple<span class="p">[</span>int<span class="p">, </span><span class="p">…</span><span class="p">]</span></em><a class="headerlink" href="#core.layers.Conv2dSame.stride" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.layers.Conv2dSame.transposed">
-<code class="sig-name descname">transposed</code><em class="property">: bool</em><a class="headerlink" href="#core.layers.Conv2dSame.transposed" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.layers.Conv2dSame.weight">
-<code class="sig-name descname">weight</code><em class="property">: torch.Tensor</em><a class="headerlink" href="#core.layers.Conv2dSame.weight" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.layers.ConvBnReluMaxPool">
-<em class="property">class </em><code class="sig-prename descclassname">core.layers.</code><code class="sig-name descname">ConvBnReluMaxPool</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">out_channels</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#ConvBnReluMaxPool"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.ConvBnReluMaxPool" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.layers.EncoderBlock" title="core.layers.EncoderBlock"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.layers.EncoderBlock</span></code></a></p>
-<p>Block of convolution, batchnorm, relu and 2x2 max pool.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>in_channels</strong> (<cite>int</cite>) – Number of input channels.</p></li>
-<li><p><strong>out_channels</strong> (<cite>int</cite>) – Number of output channels.</p></li>
-<li><p><strong>**kwargs</strong> (‘dict’ [<cite>str</cite>]) – Additional keyword arguments passed to
-<cite>pysegcnn.core.layers.Conv2dSame</cite>.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
-<td><p>Adds a child module to the current module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">apply</span></code>(fn)</p></td>
-<td><p>Applies <code class="docutils literal notranslate"><span class="pre">fn</span></code> recursively to every submodule (as returned by <code class="docutils literal notranslate"><span class="pre">.children()</span></code>) as well as self.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">bfloat16</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">bfloat16</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">buffers</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cpu</span></code>()</p></td>
-<td><p>Moves all model parameters and buffers to the CPU.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cuda</span></code>([device])</p></td>
-<td><p>Moves all model parameters and buffers to the GPU.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">double</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">double</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.layers.ConvBnReluMaxPool.downsample" title="core.layers.ConvBnReluMaxPool.downsample"><code class="xref py py-obj docutils literal notranslate"><span class="pre">downsample</span></code></a>(x)</p></td>
-<td><p>2x2 max pooling layer, <cite>torch.nn.functional.max_pool2d</cite>.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">eval</span></code>()</p></td>
-<td><p>Sets the module in evaluation mode.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.layers.ConvBnReluMaxPool.extra_repr" title="core.layers.ConvBnReluMaxPool.extra_repr"><code class="xref py py-obj docutils literal notranslate"><span class="pre">extra_repr</span></code></a>()</p></td>
-<td><p>Define optional extra information about this module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">float</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to float datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward</span></code>(x)</p></td>
-<td><p>Forward pass of an encoder block.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">half</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">half</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.layers.ConvBnReluMaxPool.layers" title="core.layers.ConvBnReluMaxPool.layers"><code class="xref py py-obj docutils literal notranslate"><span class="pre">layers</span></code></a>()</p></td>
-<td><p>Sequence of convolution, batchnorm and relu layers.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_state_dict</span></code>(state_dict[, strict])</p></td>
-<td><p>Copies parameters and buffers from <code class="xref py py-attr docutils literal notranslate"><span class="pre">state_dict</span></code> into this module and its descendants.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">modules</span></code>()</p></td>
-<td><p>Returns an iterator over all modules in the network.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_buffers</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_modules</span></code>([memo, prefix])</p></td>
-<td><p>Returns an iterator over all modules in the network, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_parameters</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parameters</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module parameters.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_backward_hook</span></code>(hook)</p></td>
-<td><p>Registers a backward hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_buffer</span></code>(name, tensor[, persistent])</p></td>
-<td><p>Adds a buffer to the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_pre_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward pre-hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_parameter</span></code>(name, param)</p></td>
-<td><p>Adds a parameter to the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">requires_grad_</span></code>([requires_grad])</p></td>
-<td><p>Change if autograd should record operations on parameters in this module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">state_dict</span></code>([destination, prefix, keep_vars])</p></td>
-<td><p>Returns a dictionary containing a whole state of the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to</span></code>(*args, **kwargs)</p></td>
-<td><p>Moves and/or casts the parameters and buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">train</span></code>([mode])</p></td>
-<td><p>Sets the module in training mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">type</span></code>(dst_type)</p></td>
-<td><p>Casts all parameters and buffers to <code class="xref py py-attr docutils literal notranslate"><span class="pre">dst_type</span></code>.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">zero_grad</span></code>()</p></td>
-<td><p>Sets gradients of all model parameters to zero.</p></td>
-</tr>
-</tbody>
-</table>
-<table class="docutils align-default">
-<colgroup>
-<col style="width: 62%" />
-<col style="width: 38%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><strong>__call__</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-even"><td><p><strong>share_memory</strong></p></td>
-<td></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="core.layers.ConvBnReluMaxPool.downsample">
-<code class="sig-name descname">downsample</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">x</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#ConvBnReluMaxPool.downsample"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.ConvBnReluMaxPool.downsample" title="Permalink to this definition">¶</a></dt>
-<dd><p>2x2 max pooling layer, <cite>torch.nn.functional.max_pool2d</cite>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>x</strong> (<cite>torch.Tensor</cite>, shape=(batch, channel, height, width)) – Input tensor.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><ul class="simple">
-<li><p><strong>x</strong> (<cite>torch.Tensor</cite>, shape=(batch, channel, height // 2, width // 2)) – The 2x2 max pooled tensor.</p></li>
-<li><p><strong>indices</strong> (<cite>torch.Tensor</cite> or <cite>None</cite>) – The indices of the maxima. Useful for upsampling with
-<cite>torch.nn.functional.max_unpool2d</cite>.</p></li>
-</ul>
-</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.layers.ConvBnReluMaxPool.extra_repr">
-<code class="sig-name descname">extra_repr</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#ConvBnReluMaxPool.extra_repr"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.ConvBnReluMaxPool.extra_repr" title="Permalink to this definition">¶</a></dt>
-<dd><p>Define optional extra information about this module.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p>Extra representation string.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>str</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.layers.ConvBnReluMaxPool.layers">
-<code class="sig-name descname">layers</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#ConvBnReluMaxPool.layers"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.ConvBnReluMaxPool.layers" title="Permalink to this definition">¶</a></dt>
-<dd><p>Sequence of convolution, batchnorm and relu layers.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>layers</strong> – An instance of <cite>torch.nn.Sequential</cite> containing the sequence
-of convolution, batchnorm and relu layer (<cite>torch.nn.Module</cite>)
-instances.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>torch.nn.Sequential</cite> [<cite>torch.nn.Module</cite>]</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt id="core.layers.ConvBnReluMaxPool.training">
-<code class="sig-name descname">training</code><em class="property">: bool</em><a class="headerlink" href="#core.layers.ConvBnReluMaxPool.training" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.layers.ConvBnReluMaxUnpool">
-<em class="property">class </em><code class="sig-prename descclassname">core.layers.</code><code class="sig-name descname">ConvBnReluMaxUnpool</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">out_channels</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#ConvBnReluMaxUnpool"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.ConvBnReluMaxUnpool" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.layers.DecoderBlock" title="core.layers.DecoderBlock"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.layers.DecoderBlock</span></code></a></p>
-<p>Block of convolution, batchnorm, relu and 2x2 max unpool.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>in_channels</strong> (<cite>int</cite>) – Number of input channels.</p></li>
-<li><p><strong>out_channels</strong> (<cite>int</cite>) – Number of output channels</p></li>
-<li><p><strong>**kwargs</strong> (‘dict’ [<cite>str</cite>]) – Additional keyword arguments passed to
-<cite>pysegcnn.core.layers.Conv2dSame</cite>.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
-<td><p>Adds a child module to the current module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">apply</span></code>(fn)</p></td>
-<td><p>Applies <code class="docutils literal notranslate"><span class="pre">fn</span></code> recursively to every submodule (as returned by <code class="docutils literal notranslate"><span class="pre">.children()</span></code>) as well as self.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">bfloat16</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">bfloat16</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">buffers</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cpu</span></code>()</p></td>
-<td><p>Moves all model parameters and buffers to the CPU.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cuda</span></code>([device])</p></td>
-<td><p>Moves all model parameters and buffers to the GPU.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">double</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">double</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">eval</span></code>()</p></td>
-<td><p>Sets the module in evaluation mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.layers.ConvBnReluMaxUnpool.extra_repr" title="core.layers.ConvBnReluMaxUnpool.extra_repr"><code class="xref py py-obj docutils literal notranslate"><span class="pre">extra_repr</span></code></a>()</p></td>
-<td><p>Define optional extra information about this module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">float</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to float datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward</span></code>(x, feature, indices, skip)</p></td>
-<td><p>Forward pass of a decoder block.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">half</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">half</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.layers.ConvBnReluMaxUnpool.layers" title="core.layers.ConvBnReluMaxUnpool.layers"><code class="xref py py-obj docutils literal notranslate"><span class="pre">layers</span></code></a>()</p></td>
-<td><p>Sequence of convolution, batchnorm and relu layers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_state_dict</span></code>(state_dict[, strict])</p></td>
-<td><p>Copies parameters and buffers from <code class="xref py py-attr docutils literal notranslate"><span class="pre">state_dict</span></code> into this module and its descendants.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">modules</span></code>()</p></td>
-<td><p>Returns an iterator over all modules in the network.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_buffers</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_modules</span></code>([memo, prefix])</p></td>
-<td><p>Returns an iterator over all modules in the network, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_parameters</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parameters</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module parameters.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_backward_hook</span></code>(hook)</p></td>
-<td><p>Registers a backward hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_buffer</span></code>(name, tensor[, persistent])</p></td>
-<td><p>Adds a buffer to the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_pre_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward pre-hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_parameter</span></code>(name, param)</p></td>
-<td><p>Adds a parameter to the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">requires_grad_</span></code>([requires_grad])</p></td>
-<td><p>Change if autograd should record operations on parameters in this module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">state_dict</span></code>([destination, prefix, keep_vars])</p></td>
-<td><p>Returns a dictionary containing a whole state of the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to</span></code>(*args, **kwargs)</p></td>
-<td><p>Moves and/or casts the parameters and buffers.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">train</span></code>([mode])</p></td>
-<td><p>Sets the module in training mode.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">type</span></code>(dst_type)</p></td>
-<td><p>Casts all parameters and buffers to <code class="xref py py-attr docutils literal notranslate"><span class="pre">dst_type</span></code>.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.layers.ConvBnReluMaxUnpool.upsample" title="core.layers.ConvBnReluMaxUnpool.upsample"><code class="xref py py-obj docutils literal notranslate"><span class="pre">upsample</span></code></a>(x, feature, indices)</p></td>
-<td><p>2x2 max unpooling layer.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">zero_grad</span></code>()</p></td>
-<td><p>Sets gradients of all model parameters to zero.</p></td>
-</tr>
-</tbody>
-</table>
-<table class="docutils align-default">
-<colgroup>
-<col style="width: 62%" />
-<col style="width: 38%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><strong>__call__</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-even"><td><p><strong>share_memory</strong></p></td>
-<td></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="core.layers.ConvBnReluMaxUnpool.extra_repr">
-<code class="sig-name descname">extra_repr</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#ConvBnReluMaxUnpool.extra_repr"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.ConvBnReluMaxUnpool.extra_repr" title="Permalink to this definition">¶</a></dt>
-<dd><p>Define optional extra information about this module.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p>Extra representation string.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>str</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.layers.ConvBnReluMaxUnpool.layers">
-<code class="sig-name descname">layers</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#ConvBnReluMaxUnpool.layers"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.ConvBnReluMaxUnpool.layers" title="Permalink to this definition">¶</a></dt>
-<dd><p>Sequence of convolution, batchnorm and relu layers.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>layers</strong> – An instance of <cite>torch.nn.Sequential</cite> containing the sequence
-of convolution, batchnorm and relu layer (<cite>torch.nn.Module</cite>)
-instances.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>torch.nn.Sequential</cite> [<cite>torch.nn.Module</cite>]</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt id="core.layers.ConvBnReluMaxUnpool.training">
-<code class="sig-name descname">training</code><em class="property">: bool</em><a class="headerlink" href="#core.layers.ConvBnReluMaxUnpool.training" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py method">
-<dt id="core.layers.ConvBnReluMaxUnpool.upsample">
-<code class="sig-name descname">upsample</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">x</span></em>, <em class="sig-param"><span class="n">feature</span></em>, <em class="sig-param"><span class="n">indices</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#ConvBnReluMaxUnpool.upsample"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.ConvBnReluMaxUnpool.upsample" title="Permalink to this definition">¶</a></dt>
-<dd><p>2x2 max unpooling layer.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>x</strong> (<cite>torch.Tensor</cite>, shape=(batch, channel, height, width)) – Input tensor.</p></li>
-<li><p><strong>feature</strong> (<cite>torch.Tensor</cite>, shape=(batch, channel, height, width)) – Intermediate output of a layer in the encoder. Used to determine
-the output shape of the upsampling operation.</p></li>
-<li><p><strong>indices</strong> (<cite>torch.Tensor</cite>) – The indices of the maxima of the max pooling operation
-(as returned by <cite>torch.nn.functional.max_pool2d</cite>).</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>x</strong> – The 2x2 max unpooled tensor.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>torch.Tensor</cite>, shape=(batch, channel, height * 2, width * 2)</p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.layers.ConvBnReluUpsample">
-<em class="property">class </em><code class="sig-prename descclassname">core.layers.</code><code class="sig-name descname">ConvBnReluUpsample</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">out_channels</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#ConvBnReluUpsample"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.ConvBnReluUpsample" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.layers.DecoderBlock" title="core.layers.DecoderBlock"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.layers.DecoderBlock</span></code></a></p>
-<p>Block of convolution, batchnorm, relu and nearest neighbor upsampling.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>in_channels</strong> (<cite>int</cite>) – Number of input channels.</p></li>
-<li><p><strong>out_channels</strong> (<cite>int</cite>) – Number of output channels</p></li>
-<li><p><strong>**kwargs</strong> (‘dict’ [<cite>str</cite>]) – Additional arguments passed to <cite>pysegcnn.core.layers.Conv2dSame</cite>.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
-<td><p>Adds a child module to the current module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">apply</span></code>(fn)</p></td>
-<td><p>Applies <code class="docutils literal notranslate"><span class="pre">fn</span></code> recursively to every submodule (as returned by <code class="docutils literal notranslate"><span class="pre">.children()</span></code>) as well as self.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">bfloat16</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">bfloat16</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">buffers</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cpu</span></code>()</p></td>
-<td><p>Moves all model parameters and buffers to the CPU.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cuda</span></code>([device])</p></td>
-<td><p>Moves all model parameters and buffers to the GPU.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">double</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">double</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">eval</span></code>()</p></td>
-<td><p>Sets the module in evaluation mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.layers.ConvBnReluUpsample.extra_repr" title="core.layers.ConvBnReluUpsample.extra_repr"><code class="xref py py-obj docutils literal notranslate"><span class="pre">extra_repr</span></code></a>()</p></td>
-<td><p>Define optional extra information about this module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">float</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to float datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward</span></code>(x, feature, indices, skip)</p></td>
-<td><p>Forward pass of a decoder block.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">half</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">half</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.layers.ConvBnReluUpsample.layers" title="core.layers.ConvBnReluUpsample.layers"><code class="xref py py-obj docutils literal notranslate"><span class="pre">layers</span></code></a>()</p></td>
-<td><p>Sequence of convolution, batchnorm and relu layers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_state_dict</span></code>(state_dict[, strict])</p></td>
-<td><p>Copies parameters and buffers from <code class="xref py py-attr docutils literal notranslate"><span class="pre">state_dict</span></code> into this module and its descendants.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">modules</span></code>()</p></td>
-<td><p>Returns an iterator over all modules in the network.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_buffers</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_modules</span></code>([memo, prefix])</p></td>
-<td><p>Returns an iterator over all modules in the network, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_parameters</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parameters</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module parameters.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_backward_hook</span></code>(hook)</p></td>
-<td><p>Registers a backward hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_buffer</span></code>(name, tensor[, persistent])</p></td>
-<td><p>Adds a buffer to the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_pre_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward pre-hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_parameter</span></code>(name, param)</p></td>
-<td><p>Adds a parameter to the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">requires_grad_</span></code>([requires_grad])</p></td>
-<td><p>Change if autograd should record operations on parameters in this module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">state_dict</span></code>([destination, prefix, keep_vars])</p></td>
-<td><p>Returns a dictionary containing a whole state of the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to</span></code>(*args, **kwargs)</p></td>
-<td><p>Moves and/or casts the parameters and buffers.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">train</span></code>([mode])</p></td>
-<td><p>Sets the module in training mode.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">type</span></code>(dst_type)</p></td>
-<td><p>Casts all parameters and buffers to <code class="xref py py-attr docutils literal notranslate"><span class="pre">dst_type</span></code>.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.layers.ConvBnReluUpsample.upsample" title="core.layers.ConvBnReluUpsample.upsample"><code class="xref py py-obj docutils literal notranslate"><span class="pre">upsample</span></code></a>(x, feature[, indices])</p></td>
-<td><p>Nearest neighbor upsampling.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">zero_grad</span></code>()</p></td>
-<td><p>Sets gradients of all model parameters to zero.</p></td>
-</tr>
-</tbody>
-</table>
-<table class="docutils align-default">
-<colgroup>
-<col style="width: 62%" />
-<col style="width: 38%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><strong>__call__</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-even"><td><p><strong>share_memory</strong></p></td>
-<td></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="core.layers.ConvBnReluUpsample.extra_repr">
-<code class="sig-name descname">extra_repr</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#ConvBnReluUpsample.extra_repr"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.ConvBnReluUpsample.extra_repr" title="Permalink to this definition">¶</a></dt>
-<dd><p>Define optional extra information about this module.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p>Extra representation string.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>str</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.layers.ConvBnReluUpsample.layers">
-<code class="sig-name descname">layers</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#ConvBnReluUpsample.layers"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.ConvBnReluUpsample.layers" title="Permalink to this definition">¶</a></dt>
-<dd><p>Sequence of convolution, batchnorm and relu layers.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>layers</strong> – An instance of <cite>torch.nn.Sequential</cite> containing the sequence
-of convolution, batchnorm and relu layer (<cite>torch.nn.Module</cite>)
-instances.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>torch.nn.Sequential</cite> [<cite>torch.nn.Module</cite>]</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt id="core.layers.ConvBnReluUpsample.training">
-<code class="sig-name descname">training</code><em class="property">: bool</em><a class="headerlink" href="#core.layers.ConvBnReluUpsample.training" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py method">
-<dt id="core.layers.ConvBnReluUpsample.upsample">
-<code class="sig-name descname">upsample</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">x</span></em>, <em class="sig-param"><span class="n">feature</span></em>, <em class="sig-param"><span class="n">indices</span><span class="o">=</span><span class="default_value">None</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#ConvBnReluUpsample.upsample"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.ConvBnReluUpsample.upsample" title="Permalink to this definition">¶</a></dt>
-<dd><p>Nearest neighbor upsampling.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>x</strong> (<cite>torch.Tensor</cite>, shape=(batch, channel, height, width)) – Input tensor.</p></li>
-<li><p><strong>feature</strong> (<cite>torch.Tensor</cite>, shape=(batch, channel, height, width)) – Intermediate output of a layer in the encoder. Used to determine
-the output shape of the upsampling operation.</p></li>
-<li><p><strong>indices</strong> (<cite>None</cite>, optional) – The indices of the maxima of the max pooling operation
-(as returned by <cite>torch.nn.functional.max_pool2d</cite>). Not required by
-this upsampling method.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>x</strong> – The 2x2 upsampled tensor.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>torch.Tensor</cite>, shape=(batch, channel, height, width)</p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.layers.Decoder">
-<em class="property">class </em><code class="sig-prename descclassname">core.layers.</code><code class="sig-name descname">Decoder</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">filters</span></em>, <em class="sig-param"><span class="n">block</span></em>, <em class="sig-param"><span class="n">skip</span><span class="o">=</span><span class="default_value">True</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#Decoder"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.Decoder" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">torch.nn.modules.module.Module</span></code></p>
-<p>Generic convolutional decoder.</p>
-<p>When instanciating an encoder-decoder architechure, <code class="docutils literal notranslate"><span class="pre">filters</span></code> should be
-the same for <cite>pysegcnn.core.layers.Encoder</cite> and
-<cite>pysegcnn.core.layers.Decoder</cite>.</p>
-<p>See <cite>pysegcnn.core.models.UNet</cite> for an example implementation.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>filters</strong> (<cite>list</cite> [<cite>int</cite>]) – List of input channels to each convolutional block. The length of
-<code class="docutils literal notranslate"><span class="pre">filters</span></code> determines the depth of the decoder. The first element of
-<code class="docutils literal notranslate"><span class="pre">filters</span></code> has to be the number of channels of the input images.</p></li>
-<li><p><strong>block</strong> (<cite>pysegcnn.core.layers.DecoderBlock</cite>) – The convolutional block defining a layer in the decoder.
-A subclass of <cite>pysegcnn.core.layers.DecoderBlock</cite>, e.g.
-<cite>pysegcnn.core.layers.ConvBnReluMaxUnpool</cite>.</p></li>
-<li><p><strong>skip</strong> (<cite>bool</cite>) – Whether to apply skip connections from the encoder to the decoder.</p></li>
-<li><p><strong>**kwargs</strong> (‘dict’ [<cite>str</cite>]) – Additional arguments passed to <cite>pysegcnn.core.layers.Conv2dSame</cite>.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
-<td><p>Adds a child module to the current module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">apply</span></code>(fn)</p></td>
-<td><p>Applies <code class="docutils literal notranslate"><span class="pre">fn</span></code> recursively to every submodule (as returned by <code class="docutils literal notranslate"><span class="pre">.children()</span></code>) as well as self.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">bfloat16</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">bfloat16</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">buffers</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cpu</span></code>()</p></td>
-<td><p>Moves all model parameters and buffers to the CPU.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cuda</span></code>([device])</p></td>
-<td><p>Moves all model parameters and buffers to the GPU.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">double</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">double</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">eval</span></code>()</p></td>
-<td><p>Sets the module in evaluation mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">extra_repr</span></code>()</p></td>
-<td><p>Set the extra representation of the module</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">float</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to float datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.layers.Decoder.forward" title="core.layers.Decoder.forward"><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward</span></code></a>(x, enc_cache)</p></td>
-<td><p>Forward pass of the decoder.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">half</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">half</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_state_dict</span></code>(state_dict[, strict])</p></td>
-<td><p>Copies parameters and buffers from <code class="xref py py-attr docutils literal notranslate"><span class="pre">state_dict</span></code> into this module and its descendants.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">modules</span></code>()</p></td>
-<td><p>Returns an iterator over all modules in the network.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_buffers</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_modules</span></code>([memo, prefix])</p></td>
-<td><p>Returns an iterator over all modules in the network, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_parameters</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parameters</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module parameters.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_backward_hook</span></code>(hook)</p></td>
-<td><p>Registers a backward hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_buffer</span></code>(name, tensor[, persistent])</p></td>
-<td><p>Adds a buffer to the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_pre_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward pre-hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_parameter</span></code>(name, param)</p></td>
-<td><p>Adds a parameter to the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">requires_grad_</span></code>([requires_grad])</p></td>
-<td><p>Change if autograd should record operations on parameters in this module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">state_dict</span></code>([destination, prefix, keep_vars])</p></td>
-<td><p>Returns a dictionary containing a whole state of the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to</span></code>(*args, **kwargs)</p></td>
-<td><p>Moves and/or casts the parameters and buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">train</span></code>([mode])</p></td>
-<td><p>Sets the module in training mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">type</span></code>(dst_type)</p></td>
-<td><p>Casts all parameters and buffers to <code class="xref py py-attr docutils literal notranslate"><span class="pre">dst_type</span></code>.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">zero_grad</span></code>()</p></td>
-<td><p>Sets gradients of all model parameters to zero.</p></td>
-</tr>
-</tbody>
-</table>
-<table class="docutils align-default">
-<colgroup>
-<col style="width: 62%" />
-<col style="width: 38%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><strong>__call__</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-even"><td><p><strong>share_memory</strong></p></td>
-<td></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="core.layers.Decoder.forward">
-<code class="sig-name descname">forward</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">x</span></em>, <em class="sig-param"><span class="n">enc_cache</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#Decoder.forward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.Decoder.forward" title="Permalink to this definition">¶</a></dt>
-<dd><p>Forward pass of the decoder.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>x</strong> (<cite>torch.Tensor</cite>, shape=(batch, channel, height, width)) – Output of the encoder.</p></li>
-<li><p><strong>enc_cache</strong> (<cite>dict</cite> [<cite>dict</cite>]) – <p>Cache dictionary. The keys of the dictionary are the number of the
-network layers and the values are dictionaries with the following
-(key, value) pairs:</p>
-<blockquote>
-<div><dl class="simple">
-<dt><code class="docutils literal notranslate"><span class="pre">"feature"</span></code></dt><dd><p>The intermediate encoder outputs (<cite>torch.Tensor</cite>).</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">"indices"</span></code></dt><dd><p>The indices of the max pooling layer (<cite>torch.Tensor</cite>).</p>
-</dd>
-</dl>
-</div></blockquote>
-</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>x</strong> – Output of the decoder.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>torch.Tensor</cite>, shape=(batch, channel, height, width)</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt id="core.layers.Decoder.training">
-<code class="sig-name descname">training</code><em class="property">: bool</em><a class="headerlink" href="#core.layers.Decoder.training" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.layers.DecoderBlock">
-<em class="property">class </em><code class="sig-prename descclassname">core.layers.</code><code class="sig-name descname">DecoderBlock</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">out_channels</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#DecoderBlock"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.DecoderBlock" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.layers.Block" title="core.layers.Block"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.layers.Block</span></code></a></p>
-<p>Block of a convolutional decoder.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>in_channels</strong> (<cite>int</cite>) – Number of input channels.</p></li>
-<li><p><strong>out_channels</strong> (<cite>int</cite>) – Number of output channels.</p></li>
-<li><p><strong>**kwargs</strong> (‘dict’ [<cite>str</cite>]) – Additional arguments passed to <cite>pysegcnn.core.layers.Conv2dSame</cite>.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
-<td><p>Adds a child module to the current module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">apply</span></code>(fn)</p></td>
-<td><p>Applies <code class="docutils literal notranslate"><span class="pre">fn</span></code> recursively to every submodule (as returned by <code class="docutils literal notranslate"><span class="pre">.children()</span></code>) as well as self.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">bfloat16</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">bfloat16</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">buffers</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cpu</span></code>()</p></td>
-<td><p>Moves all model parameters and buffers to the CPU.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cuda</span></code>([device])</p></td>
-<td><p>Moves all model parameters and buffers to the GPU.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">double</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">double</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">eval</span></code>()</p></td>
-<td><p>Sets the module in evaluation mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">extra_repr</span></code>()</p></td>
-<td><p>Set the extra representation of the module</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">float</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to float datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.layers.DecoderBlock.forward" title="core.layers.DecoderBlock.forward"><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward</span></code></a>(x, feature, indices, skip)</p></td>
-<td><p>Forward pass of a decoder block.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">half</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">half</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">layers</span></code>()</p></td>
-<td><p>Define the layers of the block.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_state_dict</span></code>(state_dict[, strict])</p></td>
-<td><p>Copies parameters and buffers from <code class="xref py py-attr docutils literal notranslate"><span class="pre">state_dict</span></code> into this module and its descendants.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">modules</span></code>()</p></td>
-<td><p>Returns an iterator over all modules in the network.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_buffers</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_modules</span></code>([memo, prefix])</p></td>
-<td><p>Returns an iterator over all modules in the network, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_parameters</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parameters</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module parameters.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_backward_hook</span></code>(hook)</p></td>
-<td><p>Registers a backward hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_buffer</span></code>(name, tensor[, persistent])</p></td>
-<td><p>Adds a buffer to the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_pre_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward pre-hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_parameter</span></code>(name, param)</p></td>
-<td><p>Adds a parameter to the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">requires_grad_</span></code>([requires_grad])</p></td>
-<td><p>Change if autograd should record operations on parameters in this module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">state_dict</span></code>([destination, prefix, keep_vars])</p></td>
-<td><p>Returns a dictionary containing a whole state of the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to</span></code>(*args, **kwargs)</p></td>
-<td><p>Moves and/or casts the parameters and buffers.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">train</span></code>([mode])</p></td>
-<td><p>Sets the module in training mode.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">type</span></code>(dst_type)</p></td>
-<td><p>Casts all parameters and buffers to <code class="xref py py-attr docutils literal notranslate"><span class="pre">dst_type</span></code>.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.layers.DecoderBlock.upsample" title="core.layers.DecoderBlock.upsample"><code class="xref py py-obj docutils literal notranslate"><span class="pre">upsample</span></code></a>(x, feature, indices)</p></td>
-<td><p>Define the upsampling method.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">zero_grad</span></code>()</p></td>
-<td><p>Sets gradients of all model parameters to zero.</p></td>
-</tr>
-</tbody>
-</table>
-<table class="docutils align-default">
-<colgroup>
-<col style="width: 62%" />
-<col style="width: 38%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><strong>__call__</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-even"><td><p><strong>share_memory</strong></p></td>
-<td></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="core.layers.DecoderBlock.forward">
-<code class="sig-name descname">forward</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">x</span></em>, <em class="sig-param"><span class="n">feature</span></em>, <em class="sig-param"><span class="n">indices</span></em>, <em class="sig-param"><span class="n">skip</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#DecoderBlock.forward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.DecoderBlock.forward" title="Permalink to this definition">¶</a></dt>
-<dd><p>Forward pass of a decoder block.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>x</strong> (<cite>torch.Tensor</cite>, shape=(batch, channel, height, width)) – Input tensor.</p></li>
-<li><p><strong>feature</strong> (<cite>torch.Tensor</cite>, shape=(batch, channel, height, width)) – Intermediate output of a layer in the encoder.
-If <code class="docutils literal notranslate"><span class="pre">skip</span></code> = True, <code class="docutils literal notranslate"><span class="pre">feature</span></code> is concatenated (along the channel
-axis) to the output of the respective upsampling layer in the
-decoder (skip connection).</p></li>
-<li><p><strong>indices</strong> (<cite>torch.Tensor</cite> or <cite>None</cite>) – Indices of the encoder downsampling method.</p></li>
-<li><p><strong>skip</strong> (<cite>bool</cite>) – Whether to apply the skip connection.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>x</strong> – Output of the decoder block.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>torch.Tensor</cite>, shape=(batch, channel, height, width)</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt id="core.layers.DecoderBlock.training">
-<code class="sig-name descname">training</code><em class="property">: bool</em><a class="headerlink" href="#core.layers.DecoderBlock.training" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py method">
-<dt id="core.layers.DecoderBlock.upsample">
-<code class="sig-name descname">upsample</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">x</span></em>, <em class="sig-param"><span class="n">feature</span></em>, <em class="sig-param"><span class="n">indices</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#DecoderBlock.upsample"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.DecoderBlock.upsample" title="Permalink to this definition">¶</a></dt>
-<dd><p>Define the upsampling method.</p>
-<p>The <cite>~pysegcnn.core.layers.DecoderBlock.upsample</cite> method should
-implement the spatial upsampling operation.</p>
-<dl class="simple">
-<dt>Use one of the following functions to upsample:</dt><dd><ul class="simple">
-<li><p><cite>torch.nn.functional.max_unpool2d</cite></p></li>
-<li><p><cite>torch.nn.functional.interpolate</cite></p></li>
-</ul>
-</dd>
-</dl>
-<p>See <cite>pysegcnn.core.layers.ConvBnReluMaxUnpool</cite> or
-<cite>pysegcnn.core.layers.ConvBnReluUpsample</cite> for an example
-implementation.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>x</strong> (<cite>torch.Tensor</cite>, shape=(batch, channel, height, width)) – Input tensor, e.g. output of a convolutional block.</p></li>
-<li><p><strong>feature</strong> (<cite>torch.Tensor</cite>, shape=(batch, channel, height, width)) – Intermediate output of a layer in the encoder. Used to implement
-skip connections.</p></li>
-<li><p><strong>indices</strong> (<cite>torch.Tensor</cite> or <cite>None</cite>) – Indices of the encoder downsampling method.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Raises</dt>
-<dd class="field-even"><p><strong>NotImplementedError</strong> – Raised if <cite>pysegcnn.core.layers.DecoderBlock</cite> is not inherited.</p>
-</dd>
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>x</strong> – The spatially upsampled tensor.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>torch.Tensor</cite>, shape=(batch, channel, height, width)</p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.layers.Encoder">
-<em class="property">class </em><code class="sig-prename descclassname">core.layers.</code><code class="sig-name descname">Encoder</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">filters</span></em>, <em class="sig-param"><span class="n">block</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#Encoder"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.Encoder" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">torch.nn.modules.module.Module</span></code></p>
-<p>Generic convolutional encoder.</p>
-<p>When instanciating an encoder-decoder architechure, <code class="docutils literal notranslate"><span class="pre">filters</span></code> should be
-the same for <cite>pysegcnn.core.layers.Encoder</cite> and
-<cite>pysegcnn.core.layers.Decoder</cite>.</p>
-<p>See <cite>pysegcnn.core.models.UNet</cite> for an example implementation.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>filters</strong> (<cite>list</cite> [<cite>int</cite>]) – List of input channels to each convolutional block. The length of
-<code class="docutils literal notranslate"><span class="pre">filters</span></code> determines the depth of the encoder. The first element of
-<code class="docutils literal notranslate"><span class="pre">filters</span></code> has to be the number of channels of the input images.</p></li>
-<li><p><strong>block</strong> (<cite>pysegcnn.core.layers.EncoderBlock</cite>) – The convolutional block defining a layer in the encoder.
-A subclass of <cite>pysegcnn.core.layers.EncoderBlock</cite>, e.g.
-<cite>pysegcnn.core.layers.ConvBnReluMaxPool</cite>.</p></li>
-<li><p><strong>**kwargs</strong> (‘dict’ [<cite>str</cite>]) – Additional arguments passed to <cite>pysegcnn.core.layers.Conv2dSame</cite>.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
-<td><p>Adds a child module to the current module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">apply</span></code>(fn)</p></td>
-<td><p>Applies <code class="docutils literal notranslate"><span class="pre">fn</span></code> recursively to every submodule (as returned by <code class="docutils literal notranslate"><span class="pre">.children()</span></code>) as well as self.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">bfloat16</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">bfloat16</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">buffers</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cpu</span></code>()</p></td>
-<td><p>Moves all model parameters and buffers to the CPU.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cuda</span></code>([device])</p></td>
-<td><p>Moves all model parameters and buffers to the GPU.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">double</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">double</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">eval</span></code>()</p></td>
-<td><p>Sets the module in evaluation mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">extra_repr</span></code>()</p></td>
-<td><p>Set the extra representation of the module</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">float</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to float datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.layers.Encoder.forward" title="core.layers.Encoder.forward"><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward</span></code></a>(x)</p></td>
-<td><p>Forward pass of the encoder.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">half</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">half</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_state_dict</span></code>(state_dict[, strict])</p></td>
-<td><p>Copies parameters and buffers from <code class="xref py py-attr docutils literal notranslate"><span class="pre">state_dict</span></code> into this module and its descendants.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">modules</span></code>()</p></td>
-<td><p>Returns an iterator over all modules in the network.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_buffers</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_modules</span></code>([memo, prefix])</p></td>
-<td><p>Returns an iterator over all modules in the network, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_parameters</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parameters</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module parameters.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_backward_hook</span></code>(hook)</p></td>
-<td><p>Registers a backward hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_buffer</span></code>(name, tensor[, persistent])</p></td>
-<td><p>Adds a buffer to the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_pre_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward pre-hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_parameter</span></code>(name, param)</p></td>
-<td><p>Adds a parameter to the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">requires_grad_</span></code>([requires_grad])</p></td>
-<td><p>Change if autograd should record operations on parameters in this module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">state_dict</span></code>([destination, prefix, keep_vars])</p></td>
-<td><p>Returns a dictionary containing a whole state of the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to</span></code>(*args, **kwargs)</p></td>
-<td><p>Moves and/or casts the parameters and buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">train</span></code>([mode])</p></td>
-<td><p>Sets the module in training mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">type</span></code>(dst_type)</p></td>
-<td><p>Casts all parameters and buffers to <code class="xref py py-attr docutils literal notranslate"><span class="pre">dst_type</span></code>.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">zero_grad</span></code>()</p></td>
-<td><p>Sets gradients of all model parameters to zero.</p></td>
-</tr>
-</tbody>
-</table>
-<table class="docutils align-default">
-<colgroup>
-<col style="width: 62%" />
-<col style="width: 38%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><strong>__call__</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-even"><td><p><strong>share_memory</strong></p></td>
-<td></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="core.layers.Encoder.forward">
-<code class="sig-name descname">forward</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">x</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#Encoder.forward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.Encoder.forward" title="Permalink to this definition">¶</a></dt>
-<dd><p>Forward pass of the encoder.</p>
-<p>Stores intermediate outputs in a dictionary. The keys of the dictionary
-are the number of the network layers and the values are dictionaries
-with the following (key, value) pairs:</p>
-<blockquote>
-<div><dl class="simple">
-<dt><code class="docutils literal notranslate"><span class="pre">"feature"</span></code></dt><dd><p>The intermediate encoder outputs (<cite>torch.Tensor</cite>).</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">"indices"</span></code></dt><dd><p>The indices of the max pooling layer, if required
-(<cite>torch.Tensor</cite>).</p>
-</dd>
-</dl>
-</div></blockquote>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>x</strong> (<cite>torch.Tensor</cite>, shape=(batch, channel, height, width)) – Input image.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>x</strong> – Output of the encoder.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>torch.Tensor</cite>, shape=(batch, channel, height, width)</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt id="core.layers.Encoder.training">
-<code class="sig-name descname">training</code><em class="property">: bool</em><a class="headerlink" href="#core.layers.Encoder.training" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.layers.EncoderBlock">
-<em class="property">class </em><code class="sig-prename descclassname">core.layers.</code><code class="sig-name descname">EncoderBlock</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">out_channels</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#EncoderBlock"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.EncoderBlock" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.layers.Block" title="core.layers.Block"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.layers.Block</span></code></a></p>
-<p>Block of a convolutional encoder.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>in_channels</strong> (<cite>int</cite>) – Number of input channels.</p></li>
-<li><p><strong>out_channels</strong> (<cite>int</cite>) – Number of output channels.</p></li>
-<li><p><strong>**kwargs</strong> (‘dict’ [<cite>str</cite>]) – Additional arguments passed to <cite>pysegcnn.core.layers.Conv2dSame</cite>.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
-<td><p>Adds a child module to the current module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">apply</span></code>(fn)</p></td>
-<td><p>Applies <code class="docutils literal notranslate"><span class="pre">fn</span></code> recursively to every submodule (as returned by <code class="docutils literal notranslate"><span class="pre">.children()</span></code>) as well as self.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">bfloat16</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">bfloat16</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">buffers</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cpu</span></code>()</p></td>
-<td><p>Moves all model parameters and buffers to the CPU.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cuda</span></code>([device])</p></td>
-<td><p>Moves all model parameters and buffers to the GPU.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">double</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">double</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.layers.EncoderBlock.downsample" title="core.layers.EncoderBlock.downsample"><code class="xref py py-obj docutils literal notranslate"><span class="pre">downsample</span></code></a>(x)</p></td>
-<td><p>Define the downsampling method.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">eval</span></code>()</p></td>
-<td><p>Sets the module in evaluation mode.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">extra_repr</span></code>()</p></td>
-<td><p>Set the extra representation of the module</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">float</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to float datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.layers.EncoderBlock.forward" title="core.layers.EncoderBlock.forward"><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward</span></code></a>(x)</p></td>
-<td><p>Forward pass of an encoder block.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">half</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">half</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">layers</span></code>()</p></td>
-<td><p>Define the layers of the block.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_state_dict</span></code>(state_dict[, strict])</p></td>
-<td><p>Copies parameters and buffers from <code class="xref py py-attr docutils literal notranslate"><span class="pre">state_dict</span></code> into this module and its descendants.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">modules</span></code>()</p></td>
-<td><p>Returns an iterator over all modules in the network.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_buffers</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_modules</span></code>([memo, prefix])</p></td>
-<td><p>Returns an iterator over all modules in the network, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_parameters</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parameters</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module parameters.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_backward_hook</span></code>(hook)</p></td>
-<td><p>Registers a backward hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_buffer</span></code>(name, tensor[, persistent])</p></td>
-<td><p>Adds a buffer to the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_pre_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward pre-hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_parameter</span></code>(name, param)</p></td>
-<td><p>Adds a parameter to the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">requires_grad_</span></code>([requires_grad])</p></td>
-<td><p>Change if autograd should record operations on parameters in this module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">state_dict</span></code>([destination, prefix, keep_vars])</p></td>
-<td><p>Returns a dictionary containing a whole state of the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to</span></code>(*args, **kwargs)</p></td>
-<td><p>Moves and/or casts the parameters and buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">train</span></code>([mode])</p></td>
-<td><p>Sets the module in training mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">type</span></code>(dst_type)</p></td>
-<td><p>Casts all parameters and buffers to <code class="xref py py-attr docutils literal notranslate"><span class="pre">dst_type</span></code>.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">zero_grad</span></code>()</p></td>
-<td><p>Sets gradients of all model parameters to zero.</p></td>
-</tr>
-</tbody>
-</table>
-<table class="docutils align-default">
-<colgroup>
-<col style="width: 62%" />
-<col style="width: 38%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><strong>__call__</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-even"><td><p><strong>share_memory</strong></p></td>
-<td></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="core.layers.EncoderBlock.downsample">
-<code class="sig-name descname">downsample</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">x</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#EncoderBlock.downsample"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.EncoderBlock.downsample" title="Permalink to this definition">¶</a></dt>
-<dd><p>Define the downsampling method.</p>
-<p>The <cite>~pysegcnn.core.layers.EncoderBlock.downsample</cite> method should
-implement the spatial pooling operation.</p>
-<dl class="simple">
-<dt>Use one of the following functions to downsample:</dt><dd><ul class="simple">
-<li><p><cite>torch.nn.functional.max_pool2d</cite></p></li>
-<li><p><cite>torch.nn.functional.interpolate</cite></p></li>
-</ul>
-</dd>
-</dl>
-<p>See <cite>pysegcnn.core.layers.ConvBnReluMaxPool</cite> for an example
-implementation.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>x</strong> (<cite>torch.Tensor</cite>, shape=(batch, channel, height, width)) – Input tensor, e.g. output of a convolutional block.</p>
-</dd>
-<dt class="field-even">Raises</dt>
-<dd class="field-even"><p><strong>NotImplementedError</strong> – Raised if <cite>pysegcnn.core.layers.EncoderBlock</cite> is not inherited.</p>
-</dd>
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><ul class="simple">
-<li><p><strong>x</strong> (<cite>torch.Tensor</cite>, shape=(batch, channel, height, width)) – The spatially downsampled tensor.</p></li>
-<li><p><strong>indices</strong> (<cite>torch.Tensor</cite> or <cite>None</cite>) – Optional indices of the downsampling method, e.g. indices of the
-maxima when using <cite>torch.nn.functional.max_pool2d</cite>. Useful for
-upsampling later. If no indices are required to upsample, simply
-return <code class="docutils literal notranslate"><span class="pre">indices</span></code> = <cite>None</cite>.</p></li>
-</ul>
-</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.layers.EncoderBlock.forward">
-<code class="sig-name descname">forward</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">x</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#EncoderBlock.forward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.EncoderBlock.forward" title="Permalink to this definition">¶</a></dt>
-<dd><p>Forward pass of an encoder block.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>x</strong> (<cite>torch.Tensor</cite>) – Input tensor, e.g. output of the previous block/layer.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><ul class="simple">
-<li><p><strong>y</strong> (<cite>torch.Tensor</cite>, shape=(batch, channel, height, width)) – Output of the encoder block.</p></li>
-<li><p><strong>x</strong> (<cite>torch.Tensor</cite>, shape=(batch, channel, height, width)) – Intermediate output before applying downsampling. Useful to
-implement skip connections.</p></li>
-<li><p><strong>indices</strong> (<cite>torch.Tensor</cite> or <cite>None</cite>) – Optional indices of the downsampling method, e.g. indices of the
-maxima when using <cite>torch.nn.functional.max_pool2d</cite>. Useful for
-upsampling later. If no indices are required to upsample, simply
-return <code class="docutils literal notranslate"><span class="pre">indices</span></code> = <cite>None</cite>.</p></li>
-</ul>
-</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt id="core.layers.EncoderBlock.training">
-<code class="sig-name descname">training</code><em class="property">: bool</em><a class="headerlink" href="#core.layers.EncoderBlock.training" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-</dd></dl>
-
-<dl class="py function">
-<dt id="core.layers.conv_bn_relu">
-<code class="sig-prename descclassname">core.layers.</code><code class="sig-name descname">conv_bn_relu</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">out_channels</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/layers.html#conv_bn_relu"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.layers.conv_bn_relu" title="Permalink to this definition">¶</a></dt>
-<dd><p>Block of convolution, batch normalization and rectified linear unit.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>in_channels</strong> (<cite>int</cite>) – Number of input channels.</p></li>
-<li><p><strong>out_channels</strong> (<cite>int</cite>) – Number of output channels.</p></li>
-<li><p><strong>**kwargs</strong> (‘dict’ [<cite>str</cite>]) – Additional arguments passed to <cite>pysegcnn.core.layers.Conv2dSame</cite>.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>block</strong> – An instance of <cite>torch.nn.Sequential</cite> containing a sequence of
-convolution, batch normalization and rectified linear unit layers.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>torch.nn.Sequential</cite> [<cite>torch.nn.Module</cite>]</p>
-</dd>
-</dl>
-</dd></dl>
-
-</div>
-</div>
-<div class="section" id="module-core.logging">
-<span id="core-logging-module"></span><h2>core.logging module<a class="headerlink" href="#module-core.logging" title="Permalink to this headline">¶</a></h2>
-<p>Logging configuration.</p>
-<div class="section" id="id6">
-<h3>License<a class="headerlink" href="#id6" title="Permalink to this headline">¶</a></h3>
-<blockquote>
-<div><p>Copyright (c) 2020 Daniel Frisinghelli</p>
-<p>This source code is licensed under the GNU General Public License v3.</p>
-<p>See the LICENSE file in the repository’s root directory.</p>
-</div></blockquote>
-<dl class="py function">
-<dt id="core.logging.log_conf">
-<code class="sig-prename descclassname">core.logging.</code><code class="sig-name descname">log_conf</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">logfile</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/logging.html#log_conf"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.logging.log_conf" title="Permalink to this definition">¶</a></dt>
-<dd><p>Set basic logging configuration passed to <cite>logging.config.dictConfig</cite>.</p>
-<p>See the logging <a class="reference external" href="https://docs.python.org/3/library/logging.config.html#dictionary-schema-details">docs</a> for a detailed description of the configuration
-dictionary.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>logfile</strong> (<cite>str</cite> or <cite>pathlib.Path</cite>) – The file to save the logs to.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>LOGGING_CONFIG</strong> – The logging configuration.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>dict</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-</div>
-</div>
-<div class="section" id="module-core.models">
-<span id="core-models-module"></span><h2>core.models module<a class="headerlink" href="#module-core.models" title="Permalink to this headline">¶</a></h2>
-<p>Neural networks for semantic image segmentation.</p>
-<div class="section" id="id7">
-<h3>License<a class="headerlink" href="#id7" title="Permalink to this headline">¶</a></h3>
-<blockquote>
-<div><p>Copyright (c) 2020 Daniel Frisinghelli</p>
-<p>This source code is licensed under the GNU General Public License v3.</p>
-<p>See the LICENSE file in the repository’s root directory.</p>
-</div></blockquote>
-<dl class="py class">
-<dt id="core.models.Network">
-<em class="property">class </em><code class="sig-prename descclassname">core.models.</code><code class="sig-name descname">Network</code><a class="reference internal" href="../_modules/core/models.html#Network"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.models.Network" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">torch.nn.modules.module.Module</span></code></p>
-<p>Generic Network class.</p>
-<p>The base class for each model. If you want to implement a new model,
-inherit the <code class="docutils literal notranslate"><span class="pre">~pysegcnn.core.models.Network</span></code> class.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p></p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p>None.</p>
-</dd>
-<dt class="field-odd">Attributes</dt>
-<dd class="field-odd"><dl class="simple">
-<dt><a class="reference internal" href="#core.models.Network.state" title="core.models.Network.state"><code class="xref py py-obj docutils literal notranslate"><span class="pre">state</span></code></a></dt><dd><p>Return the model state file.</p>
-</dd>
-</dl>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
-<td><p>Adds a child module to the current module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">apply</span></code>(fn)</p></td>
-<td><p>Applies <code class="docutils literal notranslate"><span class="pre">fn</span></code> recursively to every submodule (as returned by <code class="docutils literal notranslate"><span class="pre">.children()</span></code>) as well as self.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">bfloat16</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">bfloat16</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">buffers</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cpu</span></code>()</p></td>
-<td><p>Moves all model parameters and buffers to the CPU.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cuda</span></code>([device])</p></td>
-<td><p>Moves all model parameters and buffers to the GPU.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">double</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">double</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">eval</span></code>()</p></td>
-<td><p>Sets the module in evaluation mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">extra_repr</span></code>()</p></td>
-<td><p>Set the extra representation of the module</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">float</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to float datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward</span></code>(*input)</p></td>
-<td><p></p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.models.Network.freeze" title="core.models.Network.freeze"><code class="xref py py-obj docutils literal notranslate"><span class="pre">freeze</span></code></a>()</p></td>
-<td><p>Freeze the weights of a model.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">half</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">half</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.models.Network.load" title="core.models.Network.load"><code class="xref py py-obj docutils literal notranslate"><span class="pre">load</span></code></a>(state_file)</p></td>
-<td><p>Load a model state.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_state_dict</span></code>(state_dict[, strict])</p></td>
-<td><p>Copies parameters and buffers from <code class="xref py py-attr docutils literal notranslate"><span class="pre">state_dict</span></code> into this module and its descendants.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">modules</span></code>()</p></td>
-<td><p>Returns an iterator over all modules in the network.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_buffers</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_modules</span></code>([memo, prefix])</p></td>
-<td><p>Returns an iterator over all modules in the network, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_parameters</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parameters</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module parameters.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_backward_hook</span></code>(hook)</p></td>
-<td><p>Registers a backward hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_buffer</span></code>(name, tensor[, persistent])</p></td>
-<td><p>Adds a buffer to the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_pre_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward pre-hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_parameter</span></code>(name, param)</p></td>
-<td><p>Adds a parameter to the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">requires_grad_</span></code>([requires_grad])</p></td>
-<td><p>Change if autograd should record operations on parameters in this module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.models.Network.save" title="core.models.Network.save"><code class="xref py py-obj docutils literal notranslate"><span class="pre">save</span></code></a>(state_file, optimizer[, bands])</p></td>
-<td><p>Save the model state.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">state_dict</span></code>([destination, prefix, keep_vars])</p></td>
-<td><p>Returns a dictionary containing a whole state of the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to</span></code>(*args, **kwargs)</p></td>
-<td><p>Moves and/or casts the parameters and buffers.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">train</span></code>([mode])</p></td>
-<td><p>Sets the module in training mode.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">type</span></code>(dst_type)</p></td>
-<td><p>Casts all parameters and buffers to <code class="xref py py-attr docutils literal notranslate"><span class="pre">dst_type</span></code>.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.models.Network.unfreeze" title="core.models.Network.unfreeze"><code class="xref py py-obj docutils literal notranslate"><span class="pre">unfreeze</span></code></a>()</p></td>
-<td><p>Unfreeze the weights of a model.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">zero_grad</span></code>()</p></td>
-<td><p>Sets gradients of all model parameters to zero.</p></td>
-</tr>
-</tbody>
-</table>
-<table class="docutils align-default">
-<colgroup>
-<col style="width: 62%" />
-<col style="width: 38%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><strong>__call__</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-even"><td><p><strong>share_memory</strong></p></td>
-<td></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="core.models.Network.freeze">
-<code class="sig-name descname">freeze</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/models.html#Network.freeze"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.models.Network.freeze" title="Permalink to this definition">¶</a></dt>
-<dd><p>Freeze the weights of a model.</p>
-<p>Disables gradient computation: useful when using a pretrained model for
-inference.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p></p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p>None.</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.models.Network.load">
-<em class="property">static </em><code class="sig-name descname">load</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">state_file</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/models.html#Network.load"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.models.Network.load" title="Permalink to this definition">¶</a></dt>
-<dd><p>Load a model state.</p>
-<p>Returns the model in <code class="docutils literal notranslate"><span class="pre">state_file</span></code> with the pretrained model and
-optimizer weights. Useful when resuming training an existing model.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>state_file</strong> (<cite>str</cite> or <cite>pathlib.Path</cite>) – The model state file. Model state files are stored in
-pysegcnn/main/_models.</p>
-</dd>
-<dt class="field-even">Raises</dt>
-<dd class="field-even"><p><strong>FileNotFoundError</strong> – Raised if <code class="docutils literal notranslate"><span class="pre">state_file</span></code> does not exist.</p>
-</dd>
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><ul class="simple">
-<li><p><strong>model</strong> (<cite>pysegcnn.core.models.Network</cite>) – The pretrained model.</p></li>
-<li><p><strong>optimizer</strong> (<cite>torch.optim.Optimizer</cite>) – The optimizer used to train the model.</p></li>
-<li><p><strong>model_state</strong> (‘<cite>dict</cite>) – A dictionary containing the model and optimizer state, as
-constructed by <cite>~pysegcnn.core.Network.save</cite>.</p></li>
-</ul>
-</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.models.Network.save">
-<code class="sig-name descname">save</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">state_file</span></em>, <em class="sig-param"><span class="n">optimizer</span></em>, <em class="sig-param"><span class="n">bands</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/models.html#Network.save"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.models.Network.save" title="Permalink to this definition">¶</a></dt>
-<dd><p>Save the model state.</p>
-<p>Saves the model and optimizer states together with the model
-construction parameters, to easily re-instanciate the model.</p>
-<p>Optional <code class="docutils literal notranslate"><span class="pre">kwargs</span></code> are also saved.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>state_file</strong> (<cite>str</cite> or <cite>pathlib.Path</cite>) – Path to save the model state.</p></li>
-<li><p><strong>optimizer</strong> (<cite>torch.optim.Optimizer</cite>) – The optimizer used to train the model.</p></li>
-<li><p><strong>bands</strong> (<cite>list</cite> [<cite>str</cite>] or <cite>None</cite>, optional) – List of bands the model is trained with. The default is None.</p></li>
-<li><p><strong>**kwargs</strong> – Arbitrary keyword arguments. Each keyword argument will be saved
-as (key, value) pair in <code class="docutils literal notranslate"><span class="pre">state_file</span></code>.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>model_state</strong> – A dictionary containing the model and optimizer state</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>dict</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.models.Network.state">
-<em class="property">property </em><code class="sig-name descname">state</code><a class="headerlink" href="#core.models.Network.state" title="Permalink to this definition">¶</a></dt>
-<dd><p>Return the model state file.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>state_file</strong> – The model state file.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>pathlib.Path</cite> or <cite>None</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt id="core.models.Network.training">
-<code class="sig-name descname">training</code><em class="property">: bool</em><a class="headerlink" href="#core.models.Network.training" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py method">
-<dt id="core.models.Network.unfreeze">
-<code class="sig-name descname">unfreeze</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/models.html#Network.unfreeze"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.models.Network.unfreeze" title="Permalink to this definition">¶</a></dt>
-<dd><p>Unfreeze the weights of a model.</p>
-<p>Enables gradient computation: useful when adjusting a pretrained model
-to a new dataset.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p></p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p>None.</p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.models.SupportedLossFunctions">
-<em class="property">class </em><code class="sig-prename descclassname">core.models.</code><code class="sig-name descname">SupportedLossFunctions</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">value</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/models.html#SupportedLossFunctions"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.models.SupportedLossFunctions" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">enum.Enum</span></code></p>
-<p>Names and corresponding classes of the tested loss functions.</p>
-<dl class="py attribute">
-<dt id="core.models.SupportedLossFunctions.CrossEntropy">
-<code class="sig-name descname">CrossEntropy</code><em class="property"> = <class 'torch.nn.modules.loss.CrossEntropyLoss'></em><a class="headerlink" href="#core.models.SupportedLossFunctions.CrossEntropy" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.models.SupportedModels">
-<em class="property">class </em><code class="sig-prename descclassname">core.models.</code><code class="sig-name descname">SupportedModels</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">value</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/models.html#SupportedModels"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.models.SupportedModels" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">enum.Enum</span></code></p>
-<p>Names and corresponding classes of the implemented models.</p>
-<dl class="py attribute">
-<dt id="core.models.SupportedModels.Unet">
-<code class="sig-name descname">Unet</code><em class="property"> = <class 'core.models.UNet'></em><a class="headerlink" href="#core.models.SupportedModels.Unet" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.models.SupportedOptimizers">
-<em class="property">class </em><code class="sig-prename descclassname">core.models.</code><code class="sig-name descname">SupportedOptimizers</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">value</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/models.html#SupportedOptimizers"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.models.SupportedOptimizers" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">enum.Enum</span></code></p>
-<p>Names and corresponding classes of the tested optimizers.</p>
-<dl class="py attribute">
-<dt id="core.models.SupportedOptimizers.Adam">
-<code class="sig-name descname">Adam</code><em class="property"> = <class 'torch.optim.adam.Adam'></em><a class="headerlink" href="#core.models.SupportedOptimizers.Adam" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.models.UNet">
-<em class="property">class </em><code class="sig-prename descclassname">core.models.</code><code class="sig-name descname">UNet</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">nclasses</span></em>, <em class="sig-param"><span class="n">filters</span></em>, <em class="sig-param"><span class="n">skip</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/models.html#UNet"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.models.UNet" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.models.Network" title="core.models.Network"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.models.Network</span></code></a></p>
-<p>A PyTorch implementation of <a class="reference external" href="https://arxiv.org/abs/1505.04597">U-Net</a>.</p>
-<dl class="simple">
-<dt>Slightly modified version of U-Net:</dt><dd><ul class="simple">
-<li><p>each convolution is followed by a batch normalization layer</p></li>
-<li><p>the upsampling is implemented by a 2x2 max unpooling operation</p></li>
-</ul>
-</dd>
-</dl>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>in_channels</strong> (<cite>int</cite>) – Number of channels of the input images.</p></li>
-<li><p><strong>nclasses</strong> (<cite>int</cite>) – Number of classes.</p></li>
-<li><p><strong>filters</strong> (<cite>list</cite> [<cite>int</cite>]) – List of input channels to each convolutional block.</p></li>
-<li><p><strong>skip</strong> (<cite>bool</cite>) – Whether to apply skip connections from the encoder to the decoder.</p></li>
-<li><p><strong>**kwargs</strong> (‘dict’ [<cite>str</cite>]) – Additional keyword arguments passed to
-<cite>pysegcnn.core.layers.Conv2dSame</cite>.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-<dt class="field-even">Attributes</dt>
-<dd class="field-even"><dl class="simple">
-<dt><code class="xref py py-obj docutils literal notranslate"><span class="pre">state</span></code></dt><dd><p>Return the model state file.</p>
-</dd>
-</dl>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
-<td><p>Adds a child module to the current module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">apply</span></code>(fn)</p></td>
-<td><p>Applies <code class="docutils literal notranslate"><span class="pre">fn</span></code> recursively to every submodule (as returned by <code class="docutils literal notranslate"><span class="pre">.children()</span></code>) as well as self.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">bfloat16</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">bfloat16</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">buffers</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cpu</span></code>()</p></td>
-<td><p>Moves all model parameters and buffers to the CPU.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cuda</span></code>([device])</p></td>
-<td><p>Moves all model parameters and buffers to the GPU.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">double</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">double</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">eval</span></code>()</p></td>
-<td><p>Sets the module in evaluation mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">extra_repr</span></code>()</p></td>
-<td><p>Set the extra representation of the module</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">float</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to float datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.models.UNet.forward" title="core.models.UNet.forward"><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward</span></code></a>(x)</p></td>
-<td><p>Forward propagation of U-Net.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">freeze</span></code>()</p></td>
-<td><p>Freeze the weights of a model.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">half</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">half</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">load</span></code>(state_file)</p></td>
-<td><p>Load a model state.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_state_dict</span></code>(state_dict[, strict])</p></td>
-<td><p>Copies parameters and buffers from <code class="xref py py-attr docutils literal notranslate"><span class="pre">state_dict</span></code> into this module and its descendants.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">modules</span></code>()</p></td>
-<td><p>Returns an iterator over all modules in the network.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_buffers</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_modules</span></code>([memo, prefix])</p></td>
-<td><p>Returns an iterator over all modules in the network, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_parameters</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parameters</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module parameters.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_backward_hook</span></code>(hook)</p></td>
-<td><p>Registers a backward hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_buffer</span></code>(name, tensor[, persistent])</p></td>
-<td><p>Adds a buffer to the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_pre_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward pre-hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_parameter</span></code>(name, param)</p></td>
-<td><p>Adds a parameter to the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">requires_grad_</span></code>([requires_grad])</p></td>
-<td><p>Change if autograd should record operations on parameters in this module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">save</span></code>(state_file, optimizer[, bands])</p></td>
-<td><p>Save the model state.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">state_dict</span></code>([destination, prefix, keep_vars])</p></td>
-<td><p>Returns a dictionary containing a whole state of the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to</span></code>(*args, **kwargs)</p></td>
-<td><p>Moves and/or casts the parameters and buffers.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">train</span></code>([mode])</p></td>
-<td><p>Sets the module in training mode.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">type</span></code>(dst_type)</p></td>
-<td><p>Casts all parameters and buffers to <code class="xref py py-attr docutils literal notranslate"><span class="pre">dst_type</span></code>.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">unfreeze</span></code>()</p></td>
-<td><p>Unfreeze the weights of a model.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">zero_grad</span></code>()</p></td>
-<td><p>Sets gradients of all model parameters to zero.</p></td>
-</tr>
-</tbody>
-</table>
-<table class="docutils align-default">
-<colgroup>
-<col style="width: 62%" />
-<col style="width: 38%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><strong>__call__</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-even"><td><p><strong>share_memory</strong></p></td>
-<td></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="core.models.UNet.forward">
-<code class="sig-name descname">forward</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">x</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/models.html#UNet.forward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.models.UNet.forward" title="Permalink to this definition">¶</a></dt>
-<dd><p>Forward propagation of U-Net.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>x</strong> (<cite>torch.Tensor</cite>) – The input image, shape=(batch_size, channels, height, width).</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>y</strong> – The classified image, shape=(batch_size, height, width).</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>‘torch.tensor’</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt id="core.models.UNet.training">
-<code class="sig-name descname">training</code><em class="property">: bool</em><a class="headerlink" href="#core.models.UNet.training" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-</dd></dl>
-
-</div>
-</div>
-<div class="section" id="module-core.predict">
-<span id="core-predict-module"></span><h2>core.predict module<a class="headerlink" href="#module-core.predict" title="Permalink to this headline">¶</a></h2>
-<p>Functions for model inference.</p>
-<div class="section" id="id8">
-<h3>License<a class="headerlink" href="#id8" title="Permalink to this headline">¶</a></h3>
-<blockquote>
-<div><p>Copyright (c) 2020 Daniel Frisinghelli</p>
-<p>This source code is licensed under the GNU General Public License v3.</p>
-<p>See the LICENSE file in the repository’s root directory.</p>
-</div></blockquote>
-<dl class="py function">
-<dt id="core.predict.predict_samples">
-<code class="sig-prename descclassname">core.predict.</code><code class="sig-name descname">predict_samples</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">ds</span></em>, <em class="sig-param"><span class="n">model</span></em>, <em class="sig-param"><span class="n">cm</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">plot</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/predict.html#predict_samples"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.predict.predict_samples" title="Permalink to this definition">¶</a></dt>
-<dd><p>Classify each sample in <code class="docutils literal notranslate"><span class="pre">ds</span></code> with model <code class="docutils literal notranslate"><span class="pre">model</span></code>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>ds</strong> (<cite>pysegcnn.core.split.RandomSubset</cite> or) – </p></li>
-<li><p><strong>pysegcnn.core.split.SceneSubset</strong> – An instance of <cite>~pysegcnn.core.split.RandomSubset</cite> or
-<cite>~pysegcnn.core.split.SceneSubset</cite>.</p></li>
-<li><p><strong>model</strong> (<cite>pysegcnn.core.models.Network</cite>) – An instance of <cite>~pysegcnn.core.models.Network</cite>.</p></li>
-<li><p><strong>cm</strong> (<cite>bool</cite>, optional) – Whether to compute the confusion matrix. The default is False.</p></li>
-<li><p><strong>plot</strong> (<cite>bool</cite>, optional) – Whether to plot a false color composite, ground truth and model
-prediction for each sample. The default is False.</p></li>
-<li><p><strong>**kwargs</strong> – Additional keyword arguments passed to
-<cite>pysegcnn.core.graphics.plot_sample</cite>.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Raises</dt>
-<dd class="field-even"><p><strong>TypeError</strong> – Raised if <code class="docutils literal notranslate"><span class="pre">ds</span></code> is not an instance of
-<cite>~pysegcnn.core.split.RandomSubset</cite> or
-<cite>~pysegcnn.core.split.SceneSubset</cite>.</p>
-</dd>
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><ul>
-<li><p><strong>output</strong> (<cite>dict</cite>) –</p>
-<dl class="simple">
-<dt>Output dictionary with keys:</dt><dd><dl class="simple">
-<dt><code class="docutils literal notranslate"><span class="pre">'input'</span></code></dt><dd><p>Model input data</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'labels'</span></code></dt><dd><p>The ground truth</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'prediction'</span></code></dt><dd><p>Model prediction</p>
-</dd>
-</dl>
-</dd>
-</dl>
-</li>
-<li><p><strong>conf_mat</strong> (<cite>numpy.ndarray</cite>) – The confusion matrix. Note that the confusion matrix <code class="docutils literal notranslate"><span class="pre">conf_mat</span></code> is
-only computed if <code class="docutils literal notranslate"><span class="pre">cm</span></code> = True.</p></li>
-</ul>
-</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt id="core.predict.predict_scenes">
-<code class="sig-prename descclassname">core.predict.</code><code class="sig-name descname">predict_scenes</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">ds</span></em>, <em class="sig-param"><span class="n">model</span></em>, <em class="sig-param"><span class="n">scene_id</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="n">cm</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">plot</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/predict.html#predict_scenes"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.predict.predict_scenes" title="Permalink to this definition">¶</a></dt>
-<dd><p>Classify each scene in <code class="docutils literal notranslate"><span class="pre">ds</span></code> with model <code class="docutils literal notranslate"><span class="pre">model</span></code>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>ds</strong> (<cite>pysegcnn.core.split.SceneSubset</cite>) – An instance of <cite>~pysegcnn.core.split.SceneSubset</cite>.</p></li>
-<li><p><strong>model</strong> (<cite>pysegcnn.core.models.Network</cite>) – An instance of <cite>~pysegcnn.core.models.Network</cite>.</p></li>
-<li><p><strong>scene_id</strong> (<cite>str</cite> or <cite>None</cite>) – A valid scene identifier.</p></li>
-<li><p><strong>cm</strong> (<cite>bool</cite>, optional) – Whether to compute the confusion matrix. The default is False.</p></li>
-<li><p><strong>plot</strong> (<cite>bool</cite>, optional) – Whether to plot a false color composite, ground truth and model
-prediction for each scene. The default is False.</p></li>
-<li><p><strong>**kwargs</strong> – Additional keyword arguments passed to
-<cite>pysegcnn.core.graphics.plot_sample</cite>.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Raises</dt>
-<dd class="field-even"><p><strong>TypeError</strong> – Raised if <code class="docutils literal notranslate"><span class="pre">ds</span></code> is not an instance of
-<cite>~pysegcnn.core.split.SceneSubset</cite>.</p>
-</dd>
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><ul>
-<li><p><strong>output</strong> (<cite>dict</cite>) –</p>
-<dl class="simple">
-<dt>Output dictionary with keys:</dt><dd><dl class="simple">
-<dt><code class="docutils literal notranslate"><span class="pre">'input'</span></code></dt><dd><p>Model input data</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'labels'</span></code></dt><dd><p>The ground truth</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'prediction'</span></code></dt><dd><p>Model prediction</p>
-</dd>
-</dl>
-</dd>
-</dl>
-</li>
-<li><p><strong>conf_mat</strong> (<cite>numpy.ndarray</cite>) – The confusion matrix. Note that the confusion matrix <code class="docutils literal notranslate"><span class="pre">conf_mat</span></code> is
-only computed if <code class="docutils literal notranslate"><span class="pre">cm</span></code> = True.</p></li>
-</ul>
-</p>
-</dd>
-</dl>
-</dd></dl>
-
-</div>
-</div>
-<div class="section" id="module-core.split">
-<span id="core-split-module"></span><h2>core.split module<a class="headerlink" href="#module-core.split" title="Permalink to this headline">¶</a></h2>
-<p>Split the dataset into training, validation and test set.</p>
-<div class="section" id="id9">
-<h3>License<a class="headerlink" href="#id9" title="Permalink to this headline">¶</a></h3>
-<blockquote>
-<div><p>Copyright (c) 2020 Daniel Frisinghelli</p>
-<p>This source code is licensed under the GNU General Public License v3.</p>
-<p>See the LICENSE file in the repository’s root directory.</p>
-</div></blockquote>
-<dl class="py class">
-<dt id="core.split.CustomSubset">
-<em class="property">class </em><code class="sig-prename descclassname">core.split.</code><code class="sig-name descname">CustomSubset</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">dataset</span></em>, <em class="sig-param"><span class="n">indices</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/split.html#CustomSubset"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.split.CustomSubset" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">torch.utils.data.dataset.Subset</span></code></p>
-<p>Custom subset inheriting <cite>torch.utils.data.Subset</cite>.</p>
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.split.DateSplit">
-<em class="property">class </em><code class="sig-prename descclassname">core.split.</code><code class="sig-name descname">DateSplit</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">ds</span></em>, <em class="sig-param"><span class="n">date</span></em>, <em class="sig-param"><span class="n">dateformat</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/split.html#DateSplit"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.split.DateSplit" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.split.Split" title="core.split.Split"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.split.Split</span></code></a></p>
-<p>Split the dataset based on a date.</p>
-<p>Class wrapper for <cite>pysegcnn.core.split.Split.date_scene_split</cite>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>ds</strong> (<cite>pysegcnn.core.dataset.ImageDataset</cite>) – An instance of <cite>~pysegcnn.core.dataset.ImageDataset</cite>.</p></li>
-<li><p><strong>date</strong> (<em>'str'</em>) – A date.</p></li>
-<li><p><strong>dateformat</strong> (<em>'str'</em><em>, </em><em>optional</em>) – The format of <code class="docutils literal notranslate"><span class="pre">date</span></code>. <code class="docutils literal notranslate"><span class="pre">dateformat</span></code> is used by
-<cite>datetime.datetime.strptime’ to parse ``date`</cite> to a <cite>datetime.datetime</cite>
-object. The default is ‘%Y%m%d’.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">split</span></code>()</p></td>
-<td><p>Split dataset into training, validation and test set.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.split.DateSplit.subset_type" title="core.split.DateSplit.subset_type"><code class="xref py py-obj docutils literal notranslate"><span class="pre">subset_type</span></code></a>()</p></td>
-<td><p>Wrap <cite>pysegcnn.core.split.SceneSubset</cite>.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.split.DateSplit.subsets" title="core.split.DateSplit.subsets"><code class="xref py py-obj docutils literal notranslate"><span class="pre">subsets</span></code></a>()</p></td>
-<td><p>Wrap <cite>pysegcnn.core.split.Split.date_scene_split</cite>.</p></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="core.split.DateSplit.subset_type">
-<code class="sig-name descname">subset_type</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/split.html#DateSplit.subset_type"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.split.DateSplit.subset_type" title="Permalink to this definition">¶</a></dt>
-<dd><p>Wrap <cite>pysegcnn.core.split.SceneSubset</cite>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>SceneSubset</strong> – The subset type.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>pysegcnn.core.split.SceneSubset</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.split.DateSplit.subsets">
-<code class="sig-name descname">subsets</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/split.html#DateSplit.subsets"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.split.DateSplit.subsets" title="Permalink to this definition">¶</a></dt>
-<dd><p>Wrap <cite>pysegcnn.core.split.Split.date_scene_split</cite>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><p><strong>subsets</strong> –</p>
-<dl class="simple">
-<dt>Subset dictionary with keys:</dt><dd><dl class="simple">
-<dt><code class="docutils literal notranslate"><span class="pre">'train'</span></code></dt><dd><p>dictionary containing the training scenes.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'valid'</span></code></dt><dd><p>dictionary containing the validation scenes.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'test'</span></code></dt><dd><p>dictionary containing the test scenes, empty.</p>
-</dd>
-</dl>
-</dd>
-</dl>
-</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>dict</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.split.RandomSceneSplit">
-<em class="property">class </em><code class="sig-prename descclassname">core.split.</code><code class="sig-name descname">RandomSceneSplit</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">ds</span></em>, <em class="sig-param"><span class="n">ttratio</span></em>, <em class="sig-param"><span class="n">tvratio</span></em>, <em class="sig-param"><span class="n">seed</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/split.html#RandomSceneSplit"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.split.RandomSceneSplit" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.split.RandomSplit" title="core.split.RandomSplit"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.split.RandomSplit</span></code></a></p>
-<p>Randomly split the dataset.</p>
-<p>For each scene, all the tiles of the scene are included in either the
-training, validation or test set, respectively.</p>
-<p>Class wrapper for <cite>pysegcnn.core.split.Split.random_scene_split</cite>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>ds</strong> (<cite>pysegcnn.core.dataset.ImageDataset</cite>) – An instance of <cite>~pysegcnn.core.dataset.ImageDataset</cite>.</p></li>
-<li><p><strong>tvratio</strong> (<cite>float</cite>) – The ratio of training data to validation data, e.g. <code class="docutils literal notranslate"><span class="pre">tvratio</span></code> = 0.8
-means 80% training, 20% validation.</p></li>
-<li><p><strong>ttratio</strong> (<cite>float</cite>, optional) – The ratio of training and validation data to test data, e.g.
-<code class="docutils literal notranslate"><span class="pre">ttratio</span></code> = 0.6 means 60% for training and validation, 40% for
-testing. The default is 1.</p></li>
-<li><p><strong>seed</strong> (<cite>int</cite>, optional) – The random seed for reproducibility. The default is 0.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">split</span></code>()</p></td>
-<td><p>Split dataset into training, validation and test set.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.split.RandomSceneSplit.subset_type" title="core.split.RandomSceneSplit.subset_type"><code class="xref py py-obj docutils literal notranslate"><span class="pre">subset_type</span></code></a>()</p></td>
-<td><p>Wrap <cite>pysegcnn.core.split.SceneSubset</cite>.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.split.RandomSceneSplit.subsets" title="core.split.RandomSceneSplit.subsets"><code class="xref py py-obj docutils literal notranslate"><span class="pre">subsets</span></code></a>()</p></td>
-<td><p>Wrap <cite>pysegcnn.core.split.Split.random_scene_split</cite>.</p></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="core.split.RandomSceneSplit.subset_type">
-<code class="sig-name descname">subset_type</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/split.html#RandomSceneSplit.subset_type"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.split.RandomSceneSplit.subset_type" title="Permalink to this definition">¶</a></dt>
-<dd><p>Wrap <cite>pysegcnn.core.split.SceneSubset</cite>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>SceneSubset</strong> – The subset type.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>pysegcnn.core.split.SceneSubset</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.split.RandomSceneSplit.subsets">
-<code class="sig-name descname">subsets</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/split.html#RandomSceneSplit.subsets"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.split.RandomSceneSplit.subsets" title="Permalink to this definition">¶</a></dt>
-<dd><p>Wrap <cite>pysegcnn.core.split.Split.random_scene_split</cite>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><p><strong>subsets</strong> –</p>
-<dl class="simple">
-<dt>Subset dictionary with keys:</dt><dd><dl class="simple">
-<dt><code class="docutils literal notranslate"><span class="pre">'train'</span></code></dt><dd><p>dictionary containing the training scenes.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'valid'</span></code></dt><dd><p>dictionary containing the validation scenes.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'test'</span></code></dt><dd><p>dictionary containing the test scenes, empty.</p>
-</dd>
-</dl>
-</dd>
-</dl>
-</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>dict</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.split.RandomSplit">
-<em class="property">class </em><code class="sig-prename descclassname">core.split.</code><code class="sig-name descname">RandomSplit</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">ds</span></em>, <em class="sig-param"><span class="n">ttratio</span></em>, <em class="sig-param"><span class="n">tvratio</span></em>, <em class="sig-param"><span class="n">seed</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/split.html#RandomSplit"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.split.RandomSplit" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.split.Split" title="core.split.Split"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.split.Split</span></code></a></p>
-<p>Randomly split the dataset.</p>
-<p>Generic class for random dataset splits.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>ds</strong> (<cite>pysegcnn.core.dataset.ImageDataset</cite>) – An instance of <cite>~pysegcnn.core.dataset.ImageDataset</cite>.</p></li>
-<li><p><strong>tvratio</strong> (<cite>float</cite>) – The ratio of training data to validation data, e.g. <code class="docutils literal notranslate"><span class="pre">tvratio</span></code> = 0.8
-means 80% training, 20% validation.</p></li>
-<li><p><strong>ttratio</strong> (<cite>float</cite>, optional) – The ratio of training and validation data to test data, e.g.
-<code class="docutils literal notranslate"><span class="pre">ttratio</span></code> = 0.6 means 60% for training and validation, 40% for
-testing. The default is 1.</p></li>
-<li><p><strong>seed</strong> (<cite>int</cite>, optional) – The random seed for reproducibility. The default is 0.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">split</span></code>()</p></td>
-<td><p>Split dataset into training, validation and test set.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">subset_type</span></code>()</p></td>
-<td><p>Define the type of each subset.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">subsets</span></code>()</p></td>
-<td><p>Define training, validation and test sets.</p></td>
-</tr>
-</tbody>
-</table>
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.split.RandomSubset">
-<em class="property">class </em><code class="sig-prename descclassname">core.split.</code><code class="sig-name descname">RandomSubset</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">ds</span></em>, <em class="sig-param"><span class="n">indices</span></em>, <em class="sig-param"><span class="n">name</span></em>, <em class="sig-param"><span class="n">scenes</span></em>, <em class="sig-param"><span class="n">scene_ids</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/split.html#RandomSubset"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.split.RandomSubset" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.split.CustomSubset" title="core.split.CustomSubset"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.split.CustomSubset</span></code></a></p>
-<p>A custom subset for random dataset splits.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>ds</strong> (<cite>pysegcnn.core.dataset.ImageDataset</cite>) – An instance of <cite>~pysegcnn.core.dataset.ImageDataset</cite>.</p></li>
-<li><p><strong>indices</strong> (<cite>list</cite> [<cite>int</cite>]) – List of the subset indices to access <code class="docutils literal notranslate"><span class="pre">ds</span></code>.</p></li>
-<li><p><strong>name</strong> (<cite>str</cite>) – Name of the subset.</p></li>
-<li><p><strong>scenes</strong> (<cite>list</cite> [<cite>dict</cite>]) – List of the subset tiles.</p></li>
-<li><p><strong>scene_ids</strong> (<cite>list</cite> or <cite>numpy.ndarray</cite>) – Container of the scene ids.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.split.RandomTileSplit">
-<em class="property">class </em><code class="sig-prename descclassname">core.split.</code><code class="sig-name descname">RandomTileSplit</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">ds</span></em>, <em class="sig-param"><span class="n">ttratio</span></em>, <em class="sig-param"><span class="n">tvratio</span></em>, <em class="sig-param"><span class="n">seed</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/split.html#RandomTileSplit"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.split.RandomTileSplit" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.split.RandomSplit" title="core.split.RandomSplit"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.split.RandomSplit</span></code></a></p>
-<p>Randomly split the dataset.</p>
-<p>For each scene, the tiles of the scene can be distributed among the
-training, validation and test set.</p>
-<p>Class wrapper for <cite>pysegcnn.core.split.Split.random_tile_split</cite>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>ds</strong> (<cite>pysegcnn.core.dataset.ImageDataset</cite>) – An instance of <cite>~pysegcnn.core.dataset.ImageDataset</cite>.</p></li>
-<li><p><strong>tvratio</strong> (<cite>float</cite>) – The ratio of training data to validation data, e.g. <code class="docutils literal notranslate"><span class="pre">tvratio</span></code> = 0.8
-means 80% training, 20% validation.</p></li>
-<li><p><strong>ttratio</strong> (<cite>float</cite>, optional) – The ratio of training and validation data to test data, e.g.
-<code class="docutils literal notranslate"><span class="pre">ttratio</span></code> = 0.6 means 60% for training and validation, 40% for
-testing. The default is 1.</p></li>
-<li><p><strong>seed</strong> (<cite>int</cite>, optional) – The random seed for reproducibility. The default is 0.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">split</span></code>()</p></td>
-<td><p>Split dataset into training, validation and test set.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.split.RandomTileSplit.subset_type" title="core.split.RandomTileSplit.subset_type"><code class="xref py py-obj docutils literal notranslate"><span class="pre">subset_type</span></code></a>()</p></td>
-<td><p>Wrap <cite>pysegcnn.core.split.RandomSubset</cite>.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.split.RandomTileSplit.subsets" title="core.split.RandomTileSplit.subsets"><code class="xref py py-obj docutils literal notranslate"><span class="pre">subsets</span></code></a>()</p></td>
-<td><p>Wrap <cite>pysegcnn.core.split.Split.random_tile_split</cite>.</p></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="core.split.RandomTileSplit.subset_type">
-<code class="sig-name descname">subset_type</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/split.html#RandomTileSplit.subset_type"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.split.RandomTileSplit.subset_type" title="Permalink to this definition">¶</a></dt>
-<dd><p>Wrap <cite>pysegcnn.core.split.RandomSubset</cite>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>SceneSubset</strong> – The subset type.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>pysegcnn.core.split.RandomSubset</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.split.RandomTileSplit.subsets">
-<code class="sig-name descname">subsets</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/split.html#RandomTileSplit.subsets"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.split.RandomTileSplit.subsets" title="Permalink to this definition">¶</a></dt>
-<dd><p>Wrap <cite>pysegcnn.core.split.Split.random_tile_split</cite>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><p><strong>subsets</strong> –</p>
-<dl class="simple">
-<dt>Subset dictionary with keys:</dt><dd><dl class="simple">
-<dt><code class="docutils literal notranslate"><span class="pre">'train'</span></code></dt><dd><p>dictionary containing the training scenes.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'valid'</span></code></dt><dd><p>dictionary containing the validation scenes.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'test'</span></code></dt><dd><p>dictionary containing the test scenes, empty.</p>
-</dd>
-</dl>
-</dd>
-</dl>
-</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>dict</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.split.SceneSubset">
-<em class="property">class </em><code class="sig-prename descclassname">core.split.</code><code class="sig-name descname">SceneSubset</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">ds</span></em>, <em class="sig-param"><span class="n">indices</span></em>, <em class="sig-param"><span class="n">name</span></em>, <em class="sig-param"><span class="n">scenes</span></em>, <em class="sig-param"><span class="n">scene_ids</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/split.html#SceneSubset"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.split.SceneSubset" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.split.CustomSubset" title="core.split.CustomSubset"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.split.CustomSubset</span></code></a></p>
-<p>A custom subset for dataset splits where the scenes are preserved.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>ds</strong> (<cite>pysegcnn.core.dataset.ImageDataset</cite>) – An instance of <cite>~pysegcnn.core.dataset.ImageDataset</cite>.</p></li>
-<li><p><strong>indices</strong> (<cite>list</cite> [<cite>int</cite>]) – List of the subset indices to access <code class="docutils literal notranslate"><span class="pre">ds</span></code>.</p></li>
-<li><p><strong>name</strong> (<cite>str</cite>) – Name of the subset.</p></li>
-<li><p><strong>scenes</strong> (<cite>list</cite> [<cite>dict</cite>]) – List of the subset tiles.</p></li>
-<li><p><strong>scene_ids</strong> (<cite>list</cite> or <cite>numpy.ndarray</cite>) – Container of the scene ids.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.split.Split">
-<em class="property">class </em><code class="sig-prename descclassname">core.split.</code><code class="sig-name descname">Split</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">ds</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/split.html#Split"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.split.Split" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p>
-<p>Generic class handling how <code class="docutils literal notranslate"><span class="pre">ds</span></code> is split.</p>
-<p>Inherit <cite>~pysegcnn.core.split.Split</cite> and implement the
-<cite>~pysegcnn.core.split.Split.subsets</cite> and
-<cite>~pysegcnn.core.split.Split.subset_type</cite> method.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>ds</strong> (<cite>pysegcnn.core.dataset.ImageDataset</cite>) – An instance of <cite>~pysegcnn.core.dataset.ImageDataset</cite>.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.split.Split.split" title="core.split.Split.split"><code class="xref py py-obj docutils literal notranslate"><span class="pre">split</span></code></a>()</p></td>
-<td><p>Split dataset into training, validation and test set.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.split.Split.subset_type" title="core.split.Split.subset_type"><code class="xref py py-obj docutils literal notranslate"><span class="pre">subset_type</span></code></a>()</p></td>
-<td><p>Define the type of each subset.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.split.Split.subsets" title="core.split.Split.subsets"><code class="xref py py-obj docutils literal notranslate"><span class="pre">subsets</span></code></a>()</p></td>
-<td><p>Define training, validation and test sets.</p></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="core.split.Split.split">
-<code class="sig-name descname">split</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/split.html#Split.split"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.split.Split.split" title="Permalink to this definition">¶</a></dt>
-<dd><p>Split dataset into training, validation and test set.</p>
-<p><cite>~pysegcnn.core.split.Split.split</cite> works only if
-<cite>~pysegcnn.core.split.Split.subsets</cite> and
-<cite>~pysegcnn.core.split.Split.subset_type</cite> are implemented.</p>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.split.Split.subset_type">
-<code class="sig-name descname">subset_type</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/split.html#Split.subset_type"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.split.Split.subset_type" title="Permalink to this definition">¶</a></dt>
-<dd><p>Define the type of each subset.</p>
-<p>Wrapper method for
-<cite>pysegcnn.core.split.RandomSubset</cite> or
-<cite>pysegcnn.core.split.SceneSubset</cite>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Raises</dt>
-<dd class="field-odd"><p><strong>NotImplementedError</strong> – Raised if <cite>pysegcnn.core.split.Split</cite> is not inherited.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.split.Split.subsets">
-<code class="sig-name descname">subsets</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/split.html#Split.subsets"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.split.Split.subsets" title="Permalink to this definition">¶</a></dt>
-<dd><p>Define training, validation and test sets.</p>
-<p>Wrapper method for
-<cite>pysegcnn.core.split.Split.random_tile_split</cite>,
-<cite>pysegcnn.core.split.Split.random_scene_split</cite> or
-<cite>pysegcnn.core.split.Split.date_scene_split</cite>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Raises</dt>
-<dd class="field-odd"><p><strong>NotImplementedError</strong> – Raised if <cite>pysegcnn.core.split.Split</cite> is not inherited.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.split.SupportedSplits">
-<em class="property">class </em><code class="sig-prename descclassname">core.split.</code><code class="sig-name descname">SupportedSplits</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">value</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/split.html#SupportedSplits"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.split.SupportedSplits" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">enum.Enum</span></code></p>
-<p>Names and corresponding classes of the implemented split modes.</p>
-<dl class="py attribute">
-<dt id="core.split.SupportedSplits.date">
-<code class="sig-name descname">date</code><em class="property"> = <class 'core.split.DateSplit'></em><a class="headerlink" href="#core.split.SupportedSplits.date" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.split.SupportedSplits.random">
-<code class="sig-name descname">random</code><em class="property"> = <class 'core.split.RandomTileSplit'></em><a class="headerlink" href="#core.split.SupportedSplits.random" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.split.SupportedSplits.scene">
-<code class="sig-name descname">scene</code><em class="property"> = <class 'core.split.RandomSceneSplit'></em><a class="headerlink" href="#core.split.SupportedSplits.scene" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-</dd></dl>
-
-<dl class="py function">
-<dt id="core.split.date_scene_split">
-<code class="sig-prename descclassname">core.split.</code><code class="sig-name descname">date_scene_split</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">ds</span></em>, <em class="sig-param"><span class="n">date</span></em>, <em class="sig-param"><span class="n">dateformat</span><span class="o">=</span><span class="default_value">'%Y%m%d'</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/split.html#date_scene_split"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.split.date_scene_split" title="Permalink to this definition">¶</a></dt>
-<dd><p>Split the dataset based on a date.</p>
-<p>Scenes before <code class="docutils literal notranslate"><span class="pre">date</span></code> build the training set, scenes after <code class="docutils literal notranslate"><span class="pre">date</span></code> build
-the validation set, the test set is empty.</p>
-<p>Useful for time series data.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>ds</strong> (<cite>pysegcnn.core.dataset.ImageDataset</cite>) – An instance of <cite>~pysegcnn.core.dataset.ImageDataset</cite>.</p></li>
-<li><p><strong>date</strong> (<cite>str</cite>) – A date.</p></li>
-<li><p><strong>dateformat</strong> (<cite>str</cite>, optional) – The format of <code class="docutils literal notranslate"><span class="pre">date</span></code>. <code class="docutils literal notranslate"><span class="pre">dateformat</span></code> is used by
-<cite>datetime.datetime.strptime’ to parse ``date`</cite> to a <cite>datetime.datetime</cite>
-object. The default is ‘%Y%m%d’.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Raises</dt>
-<dd class="field-even"><p><strong>AssertionError</strong> – Raised if the splits are not pairwise disjoint.</p>
-</dd>
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><p><strong>subsets</strong> –</p>
-<dl class="simple">
-<dt>Subset dictionary with keys:</dt><dd><dl class="simple">
-<dt><code class="docutils literal notranslate"><span class="pre">'train'</span></code></dt><dd><p>dictionary containing the training scenes.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'valid'</span></code></dt><dd><p>dictionary containing the validation scenes.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'test'</span></code></dt><dd><p>dictionary containing the test scenes, empty.</p>
-</dd>
-</dl>
-</dd>
-</dl>
-</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>dict</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt id="core.split.pairwise_disjoint">
-<code class="sig-prename descclassname">core.split.</code><code class="sig-name descname">pairwise_disjoint</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">sets</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/split.html#pairwise_disjoint"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.split.pairwise_disjoint" title="Permalink to this definition">¶</a></dt>
-<dd><p>Check if <code class="docutils literal notranslate"><span class="pre">sets</span></code> are pairwise disjoint.</p>
-<p>Sets are pairwise disjoint if the length of their union equals the sum of
-their lengths.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>sets</strong> (<cite>list</cite> [<cite>collections.Sized</cite>]) – A list of sized objects.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>disjoint</strong> – Whether the sets are pairwise disjoint.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>bool</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt id="core.split.random_scene_split">
-<code class="sig-prename descclassname">core.split.</code><code class="sig-name descname">random_scene_split</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">ds</span></em>, <em class="sig-param"><span class="n">tvratio</span></em>, <em class="sig-param"><span class="n">ttratio</span><span class="o">=</span><span class="default_value">1</span></em>, <em class="sig-param"><span class="n">seed</span><span class="o">=</span><span class="default_value">0</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/split.html#random_scene_split"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.split.random_scene_split" title="Permalink to this definition">¶</a></dt>
-<dd><p>Randomly split the tiles of a dataset.</p>
-<p>For each scene, all the tiles of the scene are included in either the
-training, validation or test set, respectively.</p>
-<p>The parameters <code class="docutils literal notranslate"><span class="pre">ttratio</span></code> and <code class="docutils literal notranslate"><span class="pre">tvratio</span></code> control the size of the
-training, validation and test datasets.</p>
-<p>Test dataset size : (1 - <code class="docutils literal notranslate"><span class="pre">ttratio</span></code>) * len(<code class="docutils literal notranslate"><span class="pre">ds</span></code>)
-Train dataset size : <code class="docutils literal notranslate"><span class="pre">ttratio</span></code> * <code class="docutils literal notranslate"><span class="pre">tvratio</span></code> * len(<code class="docutils literal notranslate"><span class="pre">ds</span></code>)
-Validation dataset size: <code class="docutils literal notranslate"><span class="pre">ttratio</span></code> * (1 - <code class="docutils literal notranslate"><span class="pre">tvratio</span></code>) * len(<code class="docutils literal notranslate"><span class="pre">ds</span></code>)</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>ds</strong> (<cite>pysegcnn.core.dataset.ImageDataset</cite>) – An instance of <cite>~pysegcnn.core.dataset.ImageDataset</cite>.</p></li>
-<li><p><strong>tvratio</strong> (<cite>float</cite>) – The ratio of training data to validation data, e.g. <code class="docutils literal notranslate"><span class="pre">tvratio</span></code> = 0.8
-means 80% training, 20% validation.</p></li>
-<li><p><strong>ttratio</strong> (<cite>float</cite>, optional) – The ratio of training and validation data to test data, e.g.
-<code class="docutils literal notranslate"><span class="pre">ttratio</span></code> = 0.6 means 60% for training and validation, 40% for
-testing. The default is 1.</p></li>
-<li><p><strong>seed</strong> (<cite>int</cite>, optional) – The random seed for reproducibility. The default is 0.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Raises</dt>
-<dd class="field-even"><p><strong>AssertionError</strong> – Raised if the splits are not pairwise disjoint.</p>
-</dd>
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><p><strong>subsets</strong> –</p>
-<dl class="simple">
-<dt>Subset dictionary with keys:</dt><dd><dl class="simple">
-<dt><code class="docutils literal notranslate"><span class="pre">'train'</span></code></dt><dd><p>dictionary containing the training scenes.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'valid'</span></code></dt><dd><p>dictionary containing the validation scenes.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'test'</span></code></dt><dd><p>dictionary containing the test scenes.</p>
-</dd>
-</dl>
-</dd>
-</dl>
-</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>dict</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt id="core.split.random_tile_split">
-<code class="sig-prename descclassname">core.split.</code><code class="sig-name descname">random_tile_split</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">ds</span></em>, <em class="sig-param"><span class="n">tvratio</span></em>, <em class="sig-param"><span class="n">ttratio</span><span class="o">=</span><span class="default_value">1</span></em>, <em class="sig-param"><span class="n">seed</span><span class="o">=</span><span class="default_value">0</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/split.html#random_tile_split"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.split.random_tile_split" title="Permalink to this definition">¶</a></dt>
-<dd><p>Randomly split the tiles of a dataset.</p>
-<p>For each scene, the tiles of the scene can be distributed among the
-training, validation and test set.</p>
-<p>The parameters <code class="docutils literal notranslate"><span class="pre">ttratio</span></code> and <code class="docutils literal notranslate"><span class="pre">tvratio</span></code> control the size of the
-training, validation and test datasets.</p>
-<p>Test dataset size : (1 - <code class="docutils literal notranslate"><span class="pre">ttratio</span></code>) * len(<code class="docutils literal notranslate"><span class="pre">ds</span></code>)
-Train dataset size : <code class="docutils literal notranslate"><span class="pre">ttratio</span></code> * <code class="docutils literal notranslate"><span class="pre">tvratio</span></code> * len(<code class="docutils literal notranslate"><span class="pre">ds</span></code>)
-Validation dataset size: <code class="docutils literal notranslate"><span class="pre">ttratio</span></code> * (1 - <code class="docutils literal notranslate"><span class="pre">tvratio</span></code>) * len(<code class="docutils literal notranslate"><span class="pre">ds</span></code>)</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>ds</strong> (<cite>pysegcnn.core.dataset.ImageDataset</cite>) – An instance of <cite>~pysegcnn.core.dataset.ImageDataset</cite>.</p></li>
-<li><p><strong>tvratio</strong> (<cite>float</cite>) – The ratio of training data to validation data, e.g. <code class="docutils literal notranslate"><span class="pre">tvratio</span></code> = 0.8
-means 80% training, 20% validation.</p></li>
-<li><p><strong>ttratio</strong> (<cite>float</cite>, optional) – The ratio of training and validation data to test data, e.g.
-<code class="docutils literal notranslate"><span class="pre">ttratio</span></code> = 0.6 means 60% for training and validation, 40% for
-testing. The default is 1.</p></li>
-<li><p><strong>seed</strong> (<cite>int</cite>, optional) – The random seed for reproducibility. The default is 0.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Raises</dt>
-<dd class="field-even"><p><strong>AssertionError</strong> – Raised if the splits are not pairwise disjoint.</p>
-</dd>
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><p><strong>subsets</strong> –</p>
-<dl class="simple">
-<dt>Subset dictionary with keys:</dt><dd><dl class="simple">
-<dt><code class="docutils literal notranslate"><span class="pre">'train'</span></code></dt><dd><p>dictionary containing the training scenes.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'valid'</span></code></dt><dd><p>dictionary containing the validation scenes.</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'test'</span></code></dt><dd><p>dictionary containing the test scenes.</p>
-</dd>
-</dl>
-</dd>
-</dl>
-</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>dict</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-</div>
-</div>
-<div class="section" id="module-core.trainer">
-<span id="core-trainer-module"></span><h2>core.trainer module<a class="headerlink" href="#module-core.trainer" title="Permalink to this headline">¶</a></h2>
-<p>Model configuration and training.</p>
-<p>This module provides an end-to-end framework of dataclasses designed to train
-segmentation models on image datasets.</p>
-<p>See pysegcnn/main/train.py for a complete walkthrough.</p>
-<div class="section" id="id10">
-<h3>License<a class="headerlink" href="#id10" title="Permalink to this headline">¶</a></h3>
-<blockquote>
-<div><p>Copyright (c) 2020 Daniel Frisinghelli</p>
-<p>This source code is licensed under the GNU General Public License v3.</p>
-<p>See the LICENSE file in the repository’s root directory.</p>
-</div></blockquote>
-<dl class="py class">
-<dt id="core.trainer.BaseConfig">
-<em class="property">class </em><code class="sig-prename descclassname">core.trainer.</code><code class="sig-name descname">BaseConfig</code><a class="reference internal" href="../_modules/core/trainer.html#BaseConfig"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.BaseConfig" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p>
-<p>Base <cite>dataclasses.dataclass</cite> for each configuration.</p>
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.trainer.DatasetConfig">
-<em class="property">class </em><code class="sig-prename descclassname">core.trainer.</code><code class="sig-name descname">DatasetConfig</code><span class="sig-paren">(</span><em class="sig-param">dataset_name: str</em>, <em class="sig-param">root_dir: pathlib.Path</em>, <em class="sig-param">bands: list</em>, <em class="sig-param">tile_size: int</em>, <em class="sig-param">gt_pattern: str</em>, <em class="sig-param">seed: int</em>, <em class="sig-param">sort: bool = False</em>, <em class="sig-param">transforms: list = <factory></em>, <em class="sig-param">pad: bool = False</em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/trainer.html#DatasetConfig"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.DatasetConfig" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.trainer.BaseConfig" title="core.trainer.BaseConfig"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.trainer.BaseConfig</span></code></a></p>
-<p>Dataset configuration class.</p>
-<p>Instanciate a dataset.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>dataset_name</strong> (<cite>str</cite>) – The name of the dataset.</p></li>
-<li><p><strong>root_dir</strong> (<cite>pathlib.Path</cite>) – The root directory, path to the dataset.</p></li>
-<li><p><strong>bands</strong> (<cite>list</cite> [<cite>str</cite>]) – A list of the spectral bands to use.</p></li>
-<li><p><strong>tile_size</strong> (<cite>int</cite>) – The size of the tiles. Each scene is divided into square tiles of shape
-(tile_size, tile_size).</p></li>
-<li><p><strong>gt_pattern</strong> (<cite>str</cite>) – A pattern to match the ground truth naming convention. All directories
-and subdirectories in <code class="docutils literal notranslate"><span class="pre">root_dir</span></code> are searched for files matching
-<code class="docutils literal notranslate"><span class="pre">gt_pattern</span></code>.</p></li>
-<li><p><strong>seed</strong> (<cite>int</cite>) – The random seed. Used to split the dataset into training, validation
-and test set. Useful for reproducibility. The default is 0.</p></li>
-<li><p><strong>sort</strong> (<cite>bool</cite>, optional) – Whether to chronologically sort the samples. Useful for time series
-data. The default is False.</p></li>
-<li><p><strong>transforms</strong> (<cite>list</cite> [<cite>pysegcnn.core.split.Augment</cite>], optional) – List of <cite>pysegcnn.core.split.Augment</cite> instances. Each item in
-<code class="docutils literal notranslate"><span class="pre">transforms</span></code> generates a distinct transformed version of the dataset.
-The total dataset is composed of the original untransformed dataset
-together with each transformed version of it.
-If <code class="docutils literal notranslate"><span class="pre">transforms</span></code> = [], only the original dataset is used.
-The default is [].</p></li>
-<li><p><strong>pad</strong> (<cite>bool</cite>, optional) – Whether to center pad the input image. Set <code class="docutils literal notranslate"><span class="pre">pad</span></code> = True, if the
-images are not evenly divisible by the <code class="docutils literal notranslate"><span class="pre">tile_size</span></code>. The image data is
-padded with a constant padding value of zero. For each image, the
-corresponding ground truth image is padded with a “no data†label.
-The default is False.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.trainer.DatasetConfig.init_dataset" title="core.trainer.DatasetConfig.init_dataset"><code class="xref py py-obj docutils literal notranslate"><span class="pre">init_dataset</span></code></a>()</p></td>
-<td><p>Instanciate the dataset.</p></td>
-</tr>
-</tbody>
-</table>
-<dl class="py attribute">
-<dt id="core.trainer.DatasetConfig.bands">
-<code class="sig-name descname">bands</code><em class="property">: list</em><a class="headerlink" href="#core.trainer.DatasetConfig.bands" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.DatasetConfig.dataset_name">
-<code class="sig-name descname">dataset_name</code><em class="property">: str</em><a class="headerlink" href="#core.trainer.DatasetConfig.dataset_name" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.DatasetConfig.gt_pattern">
-<code class="sig-name descname">gt_pattern</code><em class="property">: str</em><a class="headerlink" href="#core.trainer.DatasetConfig.gt_pattern" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py method">
-<dt id="core.trainer.DatasetConfig.init_dataset">
-<code class="sig-name descname">init_dataset</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/trainer.html#DatasetConfig.init_dataset"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.DatasetConfig.init_dataset" title="Permalink to this definition">¶</a></dt>
-<dd><p>Instanciate the dataset.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>dataset</strong> – An instance of <cite>pysegcnn.core.dataset.ImageDataset</cite>.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>pysegcnn.core.dataset.ImageDataset</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.DatasetConfig.pad">
-<code class="sig-name descname">pad</code><em class="property">: bool</em><em class="property"> = False</em><a class="headerlink" href="#core.trainer.DatasetConfig.pad" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.DatasetConfig.root_dir">
-<code class="sig-name descname">root_dir</code><em class="property">: pathlib.Path</em><a class="headerlink" href="#core.trainer.DatasetConfig.root_dir" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.DatasetConfig.seed">
-<code class="sig-name descname">seed</code><em class="property">: int</em><a class="headerlink" href="#core.trainer.DatasetConfig.seed" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.DatasetConfig.sort">
-<code class="sig-name descname">sort</code><em class="property">: bool</em><em class="property"> = False</em><a class="headerlink" href="#core.trainer.DatasetConfig.sort" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.DatasetConfig.tile_size">
-<code class="sig-name descname">tile_size</code><em class="property">: int</em><a class="headerlink" href="#core.trainer.DatasetConfig.tile_size" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.DatasetConfig.transforms">
-<code class="sig-name descname">transforms</code><em class="property">: list</em><a class="headerlink" href="#core.trainer.DatasetConfig.transforms" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.trainer.EarlyStopping">
-<em class="property">class </em><code class="sig-prename descclassname">core.trainer.</code><code class="sig-name descname">EarlyStopping</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">mode</span><span class="o">=</span><span class="default_value">'max'</span></em>, <em class="sig-param"><span class="n">best</span><span class="o">=</span><span class="default_value">0</span></em>, <em class="sig-param"><span class="n">min_delta</span><span class="o">=</span><span class="default_value">0</span></em>, <em class="sig-param"><span class="n">patience</span><span class="o">=</span><span class="default_value">10</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/trainer.html#EarlyStopping"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.EarlyStopping" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p>
-<p><a href="#id25"><span class="problematic" id="id26">`Early stopping`_</span></a> algorithm.</p>
-<p>This implementation of the early stopping algorithm advances a counter each
-time a metric did not improve over a training epoch. If the metric does not
-improve over more than <code class="docutils literal notranslate"><span class="pre">patience</span></code> epochs, the early stopping criterion is
-met.</p>
-<p>See <cite>pysegcnn.core.trainer.NetworkTrainer.train</cite> for an example
-implementation.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>mode</strong> (<cite>str</cite>, optional) – The mode of the early stopping. Depends on the metric measuring
-performance. When using model loss as metric, use <code class="docutils literal notranslate"><span class="pre">mode</span></code> = ‘min’,
-however, when using accuracy as metric, use <code class="docutils literal notranslate"><span class="pre">mode</span></code> = ‘max’. The
-default is ‘max’.</p></li>
-<li><p><strong>best</strong> (<cite>float</cite>, optional) – Threshold indicating the best metric score. At instanciation, set
-<code class="docutils literal notranslate"><span class="pre">best</span></code> to the worst possible score of the metric. <code class="docutils literal notranslate"><span class="pre">best</span></code> will be
-overwritten during training. The default is 0.</p></li>
-<li><p><strong>min_delta</strong> (<cite>float</cite>, optional) – Minimum change in early stopping metric to be considered as an
-improvement. The default is 0.</p></li>
-<li><p><strong>patience</strong> (<cite>int</cite>, optional) – The number of epochs to wait for an improvement in the early stopping
-metric. The default is 10.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Raises</dt>
-<dd class="field-even"><p><strong>ValueError</strong> – Raised if <code class="docutils literal notranslate"><span class="pre">mode</span></code> is not either ‘min’ or ‘max’.</p>
-</dd>
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><ul class="simple">
-<li><p><em>None.</em></p></li>
-<li><p><em>.. _Early stopping</em> – <a class="reference external" href="https://en.wikipedia.org/wiki/Early_stopping">https://en.wikipedia.org/wiki/Early_stopping</a></p></li>
-</ul>
-</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.trainer.EarlyStopping.decreased" title="core.trainer.EarlyStopping.decreased"><code class="xref py py-obj docutils literal notranslate"><span class="pre">decreased</span></code></a>(metric, best, min_delta)</p></td>
-<td><p>Whether a metric decreased with respect to a best score.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.trainer.EarlyStopping.increased" title="core.trainer.EarlyStopping.increased"><code class="xref py py-obj docutils literal notranslate"><span class="pre">increased</span></code></a>(metric, best, min_delta)</p></td>
-<td><p>Whether a metric increased with respect to a best score.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.trainer.EarlyStopping.stop" title="core.trainer.EarlyStopping.stop"><code class="xref py py-obj docutils literal notranslate"><span class="pre">stop</span></code></a>(metric)</p></td>
-<td><p>Advance early stopping counter.</p></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="core.trainer.EarlyStopping.decreased">
-<code class="sig-name descname">decreased</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">metric</span></em>, <em class="sig-param"><span class="n">best</span></em>, <em class="sig-param"><span class="n">min_delta</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/trainer.html#EarlyStopping.decreased"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.EarlyStopping.decreased" title="Permalink to this definition">¶</a></dt>
-<dd><p>Whether a metric decreased with respect to a best score.</p>
-<p>Measure improvement for metrics that are considered as ‘better’ when
-they decrease, e.g. model loss, mean squared error, etc.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>metric</strong> (<cite>float</cite>) – The current score.</p></li>
-<li><p><strong>best</strong> (<cite>float</cite>) – The current best score.</p></li>
-<li><p><strong>min_delta</strong> (<cite>float</cite>) – Minimum change to be considered as an improvement.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p>Whether the metric improved.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>bool</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.trainer.EarlyStopping.increased">
-<code class="sig-name descname">increased</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">metric</span></em>, <em class="sig-param"><span class="n">best</span></em>, <em class="sig-param"><span class="n">min_delta</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/trainer.html#EarlyStopping.increased"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.EarlyStopping.increased" title="Permalink to this definition">¶</a></dt>
-<dd><p>Whether a metric increased with respect to a best score.</p>
-<p>Measure improvement for metrics that are considered as ‘better’ when
-they increase, e.g. accuracy, precision, recall, etc.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>metric</strong> (<cite>float</cite>) – The current score.</p></li>
-<li><p><strong>best</strong> (<cite>float</cite>) – The current best score.</p></li>
-<li><p><strong>min_delta</strong> (<cite>float</cite>) – Minimum change to be considered as an improvement.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p>Whether the metric improved.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>bool</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.trainer.EarlyStopping.stop">
-<code class="sig-name descname">stop</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">metric</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/trainer.html#EarlyStopping.stop"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.EarlyStopping.stop" title="Permalink to this definition">¶</a></dt>
-<dd><p>Advance early stopping counter.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>metric</strong> (<cite>float</cite>) – The current metric score.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>early_stop</strong> – Whether the early stopping criterion is met.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>bool</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.trainer.EvalConfig">
-<em class="property">class </em><code class="sig-prename descclassname">core.trainer.</code><code class="sig-name descname">EvalConfig</code><span class="sig-paren">(</span><em class="sig-param">state_file: pathlib.Path</em>, <em class="sig-param">test: object</em>, <em class="sig-param">predict_scene: bool = False</em>, <em class="sig-param">plot_samples: bool = False</em>, <em class="sig-param">plot_scenes: bool = False</em>, <em class="sig-param">plot_bands: list = <factory></em>, <em class="sig-param">cm: bool = True</em>, <em class="sig-param">figsize: tuple = (10</em>, <em class="sig-param">10)</em>, <em class="sig-param">alpha: int = 5</em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/trainer.html#EvalConfig"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.EvalConfig" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.trainer.BaseConfig" title="core.trainer.BaseConfig"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.trainer.BaseConfig</span></code></a></p>
-<p>Model inference configuration.</p>
-<p>Evaluate a model.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>state_file</strong> (<cite>pathlib.Path</cite>) – Path to the model to evaluate.</p></li>
-<li><p><strong>test</strong> (<cite>bool</cite> or <cite>None</cite>) – Whether to evaluate the model on the training(<code class="docutils literal notranslate"><span class="pre">test</span></code> = <cite>None</cite>), the
-validation (<code class="docutils literal notranslate"><span class="pre">test</span></code> = False) or the test set (<code class="docutils literal notranslate"><span class="pre">test</span></code> = True).</p></li>
-<li><p><strong>predict_scene</strong> (<cite>bool</cite>, optional) – The model prediction order. If False, the samples (tiles) of a dataset
-are predicted in any order and the scenes are not reconstructed.
-If True, the samples (tiles) are ordered according to the scene they
-belong to and a model prediction for each entire reconstructed scene is
-returned. The default is False.</p></li>
-<li><p><strong>plot_samples</strong> (<cite>bool</cite>, optional) – Whether to save a plot of false color composite, ground truth and model
-prediction for each sample (tile). Only used if <code class="docutils literal notranslate"><span class="pre">predict_scene</span></code> =
-False. The default is False.</p></li>
-<li><p><strong>plot_scenes</strong> (<cite>bool</cite>, optional) – Whether to save a plot of false color composite, ground truth and model
-prediction for each entire scene. Only used if <code class="docutils literal notranslate"><span class="pre">predict_scene</span></code> =
-True. The default is False.</p></li>
-<li><p><strong>plot_bands</strong> (<cite>list</cite> [<cite>str</cite>], optional) – The bands to build the false color composite. The default is
-[‘nir’, ‘red’, ‘green’].</p></li>
-<li><p><strong>cm</strong> (<cite>bool</cite>, optional) – Whether to compute and plot the confusion matrix. The default is True.</p></li>
-<li><p><strong>figsize</strong> (<cite>tuple</cite>, optional) – The figure size in centimeters. The default is (10, 10).</p></li>
-<li><p><strong>alpha</strong> (<cite>int</cite>, optional) – The level of the percentiles for contrast stretching of the false color
-compsite. The default is 0, i.e. no stretching.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<dl class="py attribute">
-<dt id="core.trainer.EvalConfig.alpha">
-<code class="sig-name descname">alpha</code><em class="property">: int</em><em class="property"> = 5</em><a class="headerlink" href="#core.trainer.EvalConfig.alpha" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.EvalConfig.cm">
-<code class="sig-name descname">cm</code><em class="property">: bool</em><em class="property"> = True</em><a class="headerlink" href="#core.trainer.EvalConfig.cm" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.EvalConfig.figsize">
-<code class="sig-name descname">figsize</code><em class="property">: tuple</em><em class="property"> = (10, 10)</em><a class="headerlink" href="#core.trainer.EvalConfig.figsize" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.EvalConfig.plot_bands">
-<code class="sig-name descname">plot_bands</code><em class="property">: list</em><a class="headerlink" href="#core.trainer.EvalConfig.plot_bands" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.EvalConfig.plot_samples">
-<code class="sig-name descname">plot_samples</code><em class="property">: bool</em><em class="property"> = False</em><a class="headerlink" href="#core.trainer.EvalConfig.plot_samples" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.EvalConfig.plot_scenes">
-<code class="sig-name descname">plot_scenes</code><em class="property">: bool</em><em class="property"> = False</em><a class="headerlink" href="#core.trainer.EvalConfig.plot_scenes" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.EvalConfig.predict_scene">
-<code class="sig-name descname">predict_scene</code><em class="property">: bool</em><em class="property"> = False</em><a class="headerlink" href="#core.trainer.EvalConfig.predict_scene" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.EvalConfig.state_file">
-<code class="sig-name descname">state_file</code><em class="property">: pathlib.Path</em><a class="headerlink" href="#core.trainer.EvalConfig.state_file" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.EvalConfig.test">
-<code class="sig-name descname">test</code><em class="property">: object</em><a class="headerlink" href="#core.trainer.EvalConfig.test" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.trainer.LogConfig">
-<em class="property">class </em><code class="sig-prename descclassname">core.trainer.</code><code class="sig-name descname">LogConfig</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">state_file</span><span class="p">:</span> <span class="n">pathlib.Path</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/trainer.html#LogConfig"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.LogConfig" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.trainer.BaseConfig" title="core.trainer.BaseConfig"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.trainer.BaseConfig</span></code></a></p>
-<p>Logging configuration class.</p>
-<p>Generate the model log file.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>state_file</strong> (<cite>pathlib.Path</cite>) – Path to a model state file.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.trainer.LogConfig.init_log" title="core.trainer.LogConfig.init_log"><code class="xref py py-obj docutils literal notranslate"><span class="pre">init_log</span></code></a>(init_str)</p></td>
-<td><p>Generate a string to identify a new model run.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.trainer.LogConfig.now" title="core.trainer.LogConfig.now"><code class="xref py py-obj docutils literal notranslate"><span class="pre">now</span></code></a>()</p></td>
-<td><p>Return the current date and time.</p></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="core.trainer.LogConfig.init_log">
-<em class="property">static </em><code class="sig-name descname">init_log</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">init_str</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/trainer.html#LogConfig.init_log"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.LogConfig.init_log" title="Permalink to this definition">¶</a></dt>
-<dd><p>Generate a string to identify a new model run.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>init_str</strong> (<cite>str</cite>) – The string to write to the model log file.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.trainer.LogConfig.now">
-<em class="property">static </em><code class="sig-name descname">now</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/trainer.html#LogConfig.now"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.LogConfig.now" title="Permalink to this definition">¶</a></dt>
-<dd><p>Return the current date and time.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>date</strong> – The current date and time.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>datetime.datetime</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.LogConfig.state_file">
-<code class="sig-name descname">state_file</code><em class="property">: pathlib.Path</em><a class="headerlink" href="#core.trainer.LogConfig.state_file" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.trainer.ModelConfig">
-<em class="property">class </em><code class="sig-prename descclassname">core.trainer.</code><code class="sig-name descname">ModelConfig</code><span class="sig-paren">(</span><em class="sig-param">model_name: str</em>, <em class="sig-param">filters: list</em>, <em class="sig-param">torch_seed: int</em>, <em class="sig-param">optim_name: str</em>, <em class="sig-param">loss_name: str</em>, <em class="sig-param">skip_connection: bool = True</em>, <em class="sig-param">kwargs: dict = <factory></em>, <em class="sig-param">batch_size: int = 64</em>, <em class="sig-param">checkpoint: bool = False</em>, <em class="sig-param">transfer: bool = False</em>, <em class="sig-param">pretrained_model: str = ''</em>, <em class="sig-param">lr: float = 0.001</em>, <em class="sig-param">early_stop: bool = False</em>, <em class="sig-param">mode: str = 'max'</em>, <em class="sig-param">delta: float = 0</em>, <em class="sig-param">patience: int = 10</em>, <em class="sig-param">epochs: int = 50</em>, <em class="sig-param">nthreads: int = 2</em>, <em class="sig-param">save: bool = True</em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/trainer.html#ModelConfig"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.ModelConfig" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.trainer.BaseConfig" title="core.trainer.BaseConfig"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.trainer.BaseConfig</span></code></a></p>
-<p>Model configuration class.</p>
-<p>Instanciate a (pretrained) model.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>model_name</strong> (<cite>str</cite>) – The name of the model.</p></li>
-<li><p><strong>filters</strong> (<cite>list</cite> [<cite>int</cite>]) – List of input channels to the convolutional layers.</p></li>
-<li><p><strong>torch_seed</strong> (<cite>int</cite>) – The random seed to initialize the model weights.
-Useful for reproducibility.</p></li>
-<li><p><strong>optim_name</strong> (<cite>str</cite>) – The name of the optimizer to update the model weights.</p></li>
-<li><p><strong>loss_name</strong> (<cite>str</cite>) – The name of the loss function measuring the model error.</p></li>
-<li><p><strong>skip_connection</strong> (<cite>bool</cite>, optional) – Whether to apply skip connections. The defaul is True.</p></li>
-<li><p><strong>kwargs</strong> (<cite>dict</cite>, optional) – The configuration for each convolution in the model. The default is
-{‘kernel_size’: 3, ‘stride’: 1, ‘dilation’: 1}.</p></li>
-<li><p><strong>batch_size</strong> (<cite>int</cite>, optional) – The model batch size. Determines the number of samples to process
-before updating the model weights. The default is 64.</p></li>
-<li><p><strong>checkpoint</strong> (<cite>bool</cite>, optional) – Whether to resume training from an existing model checkpoint. The
-default is False.</p></li>
-<li><p><strong>transfer</strong> (<cite>bool</cite>, optional) – Whether to use a model for transfer learning on a new dataset. If True,
-the model architecture of <code class="docutils literal notranslate"><span class="pre">pretrained_model</span></code> is adjusted to a new
-dataset. The default is False.</p></li>
-<li><p><strong>pretrained_model</strong> (<cite>str</cite>, optional) – The name of the pretrained model to use for transfer learning.
-The default is ‘’.</p></li>
-<li><p><strong>lr</strong> (<cite>float</cite>, optional) – The learning rate used by the gradient descent algorithm.
-The default is 0.001.</p></li>
-<li><p><strong>early_stop</strong> (<cite>bool</cite>, optional) – Whether to apply <a href="#id27"><span class="problematic" id="id28">`early stopping`_</span></a>. The default is False.</p></li>
-<li><p><strong>mode</strong> (<cite>str</cite>, optional) – The mode of the early stopping. Depends on the metric measuring
-performance. When using model loss as metric, use <code class="docutils literal notranslate"><span class="pre">mode</span></code> = ‘min’,
-however, when using accuracy as metric, use <code class="docutils literal notranslate"><span class="pre">mode</span></code> = ‘max’. For now,
-only <code class="docutils literal notranslate"><span class="pre">mode</span></code> = ‘max’ is supported. Only used if <code class="docutils literal notranslate"><span class="pre">early_stop</span></code> = True.
-The default is ‘max’.</p></li>
-<li><p><strong>delta</strong> (<cite>float</cite>, optional) – Minimum change in early stopping metric to be considered as an
-improvement. Only used if <code class="docutils literal notranslate"><span class="pre">early_stop</span></code> = True. The default is 0.</p></li>
-<li><p><strong>patience</strong> (<cite>int</cite>, optional) – The number of epochs to wait for an improvement in the early stopping
-metric. If the model does not improve over more than <code class="docutils literal notranslate"><span class="pre">patience</span></code>
-epochs, quit training. Only used if <code class="docutils literal notranslate"><span class="pre">early_stop</span></code> = True.
-The default is 10.</p></li>
-<li><p><strong>epochs</strong> (<cite>int</cite>, optional) – The maximum number of epochs to train. The default is 50.</p></li>
-<li><p><strong>nthreads</strong> (<cite>int</cite>, optional) – The number of cpu threads to use during training. The default is
-torch.get_num_threads().</p></li>
-<li><p><strong>save</strong> (<cite>bool</cite>, optional) – Whether to save the model state to disk. Model states are saved in
-pysegcnn/main/_models. The default is True.</p></li>
-<li><p><strong>_early stopping</strong> (<em>.</em>) – <a class="reference external" href="https://en.wikipedia.org/wiki/Early_stopping">https://en.wikipedia.org/wiki/Early_stopping</a></p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.trainer.ModelConfig.init_loss_function" title="core.trainer.ModelConfig.init_loss_function"><code class="xref py py-obj docutils literal notranslate"><span class="pre">init_loss_function</span></code></a>()</p></td>
-<td><p>Instanciate the loss function.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.trainer.ModelConfig.init_model" title="core.trainer.ModelConfig.init_model"><code class="xref py py-obj docutils literal notranslate"><span class="pre">init_model</span></code></a>(ds, state_file)</p></td>
-<td><p>Instanciate the model and the optimizer.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.trainer.ModelConfig.init_optimizer" title="core.trainer.ModelConfig.init_optimizer"><code class="xref py py-obj docutils literal notranslate"><span class="pre">init_optimizer</span></code></a>(model)</p></td>
-<td><p>Instanciate the optimizer.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.trainer.ModelConfig.load_checkpoint" title="core.trainer.ModelConfig.load_checkpoint"><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_checkpoint</span></code></a>(model, optimizer, state_file)</p></td>
-<td><p>Load an existing model checkpoint.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.trainer.ModelConfig.transfer_model" title="core.trainer.ModelConfig.transfer_model"><code class="xref py py-obj docutils literal notranslate"><span class="pre">transfer_model</span></code></a>(state_file, ds)</p></td>
-<td><p>Adjust a pretrained model to a new dataset.</p></td>
-</tr>
-</tbody>
-</table>
-<dl class="py attribute">
-<dt id="core.trainer.ModelConfig.batch_size">
-<code class="sig-name descname">batch_size</code><em class="property">: int</em><em class="property"> = 64</em><a class="headerlink" href="#core.trainer.ModelConfig.batch_size" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.ModelConfig.checkpoint">
-<code class="sig-name descname">checkpoint</code><em class="property">: bool</em><em class="property"> = False</em><a class="headerlink" href="#core.trainer.ModelConfig.checkpoint" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.ModelConfig.delta">
-<code class="sig-name descname">delta</code><em class="property">: float</em><em class="property"> = 0</em><a class="headerlink" href="#core.trainer.ModelConfig.delta" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.ModelConfig.early_stop">
-<code class="sig-name descname">early_stop</code><em class="property">: bool</em><em class="property"> = False</em><a class="headerlink" href="#core.trainer.ModelConfig.early_stop" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.ModelConfig.epochs">
-<code class="sig-name descname">epochs</code><em class="property">: int</em><em class="property"> = 50</em><a class="headerlink" href="#core.trainer.ModelConfig.epochs" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.ModelConfig.filters">
-<code class="sig-name descname">filters</code><em class="property">: list</em><a class="headerlink" href="#core.trainer.ModelConfig.filters" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py method">
-<dt id="core.trainer.ModelConfig.init_loss_function">
-<code class="sig-name descname">init_loss_function</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/trainer.html#ModelConfig.init_loss_function"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.ModelConfig.init_loss_function" title="Permalink to this definition">¶</a></dt>
-<dd><p>Instanciate the loss function.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>loss_function</strong> – An instance of <cite>torch.nn.Module</cite>.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>torch.nn.Module</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.trainer.ModelConfig.init_model">
-<code class="sig-name descname">init_model</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">ds</span></em>, <em class="sig-param"><span class="n">state_file</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/trainer.html#ModelConfig.init_model"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.ModelConfig.init_model" title="Permalink to this definition">¶</a></dt>
-<dd><p>Instanciate the model and the optimizer.</p>
-<p>If the model checkpoint <code class="docutils literal notranslate"><span class="pre">state_file</span></code> exists, the pretrained model and
-optimizer states are loaded, otherwise the model and the optimizer are
-initialized from scratch.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>ds</strong> (<cite>pysegcnn.core.dataset.ImageDataset</cite>) – An instance of <cite>pysegcnn.core.dataset.ImageDataset</cite>.</p></li>
-<li><p><strong>state_file</strong> (<cite>pathlib.Path</cite>) – Path to a model checkpoint.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><ul>
-<li><p><strong>model</strong> (<cite>pysegcnn.core.models.Network</cite>) – An instance of <cite>pysegcnn.core.models.Network</cite>.</p></li>
-<li><p><strong>optimizer</strong> (<cite>torch.optim.Optimizer</cite>) – An instance of <cite>torch.optim.Optimizer</cite>.</p></li>
-<li><p><strong>checkpoint_state</strong> (<cite>dict</cite> [<cite>str</cite>, <cite>numpy.ndarray</cite>]) – If the model checkpoint <code class="docutils literal notranslate"><span class="pre">state_file</span></code> exists, <code class="docutils literal notranslate"><span class="pre">checkpoint_state</span></code>
-has keys:</p>
-<blockquote>
-<div><dl class="simple">
-<dt><code class="docutils literal notranslate"><span class="pre">'ta'</span></code></dt><dd><p>The accuracy on the training set (<cite>numpy.ndarray</cite>).</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'tl'</span></code></dt><dd><p>The loss on the training set (<cite>numpy.ndarray</cite>).</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'va'</span></code></dt><dd><p>The accuracy on the validation set (<cite>numpy.ndarray</cite>).</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'vl'</span></code></dt><dd><p>The loss on the validation set (<cite>numpy.ndarray</cite>).</p>
-</dd>
-</dl>
-</div></blockquote>
-</li>
-</ul>
-</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt id="core.trainer.ModelConfig.init_optimizer">
-<code class="sig-name descname">init_optimizer</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">model</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/trainer.html#ModelConfig.init_optimizer"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.ModelConfig.init_optimizer" title="Permalink to this definition">¶</a></dt>
-<dd><p>Instanciate the optimizer.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>model</strong> (<cite>torch.nn.Module</cite>) – An instance of <cite>torch.nn.Module</cite>.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>optimizer</strong> – An instance of <cite>torch.optim.Optimizer</cite>.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>torch.optim.Optimizer</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.ModelConfig.kwargs">
-<code class="sig-name descname">kwargs</code><em class="property">: dict</em><a class="headerlink" href="#core.trainer.ModelConfig.kwargs" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py method">
-<dt id="core.trainer.ModelConfig.load_checkpoint">
-<em class="property">static </em><code class="sig-name descname">load_checkpoint</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">model</span></em>, <em class="sig-param"><span class="n">optimizer</span></em>, <em class="sig-param"><span class="n">state_file</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/trainer.html#ModelConfig.load_checkpoint"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.ModelConfig.load_checkpoint" title="Permalink to this definition">¶</a></dt>
-<dd><p>Load an existing model checkpoint.</p>
-<p>If the model checkpoint <code class="docutils literal notranslate"><span class="pre">state_file</span></code> exists, the pretrained model and
-optimizer states are loaded.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>model</strong> (<cite>pysegcnn.core.models.Network</cite>) – An instance of <cite>pysegcnn.core.models.Network</cite>.</p></li>
-<li><p><strong>optimizer</strong> (<cite>torch.optim.Optimizer</cite>) – An instance of <cite>torch.optim.Optimizer</cite>.</p></li>
-<li><p><strong>state_file</strong> (<cite>pathlib.Path</cite>) – Path to the model checkpoint.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><ul>
-<li><p><strong>model</strong> (<cite>pysegcnn.core.models.Network</cite>) – An instance of <cite>pysegcnn.core.models.Network</cite>.</p></li>
-<li><p><strong>optimizer</strong> (<cite>torch.optim.Optimizer</cite>) – An instance of <cite>torch.optim.Optimizer</cite>.</p></li>
-<li><p><strong>checkpoint_state</strong> (<cite>dict</cite> [<cite>str</cite>, <cite>numpy.ndarray</cite>]) – If the model checkpoint <code class="docutils literal notranslate"><span class="pre">state_file</span></code> exists, <code class="docutils literal notranslate"><span class="pre">checkpoint_state</span></code>
-has keys:</p>
-<blockquote>
-<div><dl class="simple">
-<dt><code class="docutils literal notranslate"><span class="pre">'ta'</span></code></dt><dd><p>The accuracy on the training set (<cite>numpy.ndarray</cite>).</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'tl'</span></code></dt><dd><p>The loss on the training set (<cite>numpy.ndarray</cite>).</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'va'</span></code></dt><dd><p>The accuracy on the validation set (<cite>numpy.ndarray</cite>).</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'vl'</span></code></dt><dd><p>The loss on the validation set (<cite>numpy.ndarray</cite>).</p>
-</dd>
-</dl>
-</div></blockquote>
-</li>
-</ul>
-</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.ModelConfig.loss_name">
-<code class="sig-name descname">loss_name</code><em class="property">: str</em><a class="headerlink" href="#core.trainer.ModelConfig.loss_name" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.ModelConfig.lr">
-<code class="sig-name descname">lr</code><em class="property">: float</em><em class="property"> = 0.001</em><a class="headerlink" href="#core.trainer.ModelConfig.lr" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.ModelConfig.mode">
-<code class="sig-name descname">mode</code><em class="property">: str</em><em class="property"> = 'max'</em><a class="headerlink" href="#core.trainer.ModelConfig.mode" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.ModelConfig.model_name">
-<code class="sig-name descname">model_name</code><em class="property">: str</em><a class="headerlink" href="#core.trainer.ModelConfig.model_name" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.ModelConfig.nthreads">
-<code class="sig-name descname">nthreads</code><em class="property">: int</em><em class="property"> = 2</em><a class="headerlink" href="#core.trainer.ModelConfig.nthreads" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.ModelConfig.optim_name">
-<code class="sig-name descname">optim_name</code><em class="property">: str</em><a class="headerlink" href="#core.trainer.ModelConfig.optim_name" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.ModelConfig.patience">
-<code class="sig-name descname">patience</code><em class="property">: int</em><em class="property"> = 10</em><a class="headerlink" href="#core.trainer.ModelConfig.patience" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.ModelConfig.pretrained_model">
-<code class="sig-name descname">pretrained_model</code><em class="property">: str</em><em class="property"> = ''</em><a class="headerlink" href="#core.trainer.ModelConfig.pretrained_model" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.ModelConfig.save">
-<code class="sig-name descname">save</code><em class="property">: bool</em><em class="property"> = True</em><a class="headerlink" href="#core.trainer.ModelConfig.save" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.ModelConfig.skip_connection">
-<code class="sig-name descname">skip_connection</code><em class="property">: bool</em><em class="property"> = True</em><a class="headerlink" href="#core.trainer.ModelConfig.skip_connection" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.ModelConfig.torch_seed">
-<code class="sig-name descname">torch_seed</code><em class="property">: int</em><a class="headerlink" href="#core.trainer.ModelConfig.torch_seed" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.ModelConfig.transfer">
-<code class="sig-name descname">transfer</code><em class="property">: bool</em><em class="property"> = False</em><a class="headerlink" href="#core.trainer.ModelConfig.transfer" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py method">
-<dt id="core.trainer.ModelConfig.transfer_model">
-<em class="property">static </em><code class="sig-name descname">transfer_model</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">state_file</span></em>, <em class="sig-param"><span class="n">ds</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/trainer.html#ModelConfig.transfer_model"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.ModelConfig.transfer_model" title="Permalink to this definition">¶</a></dt>
-<dd><p>Adjust a pretrained model to a new dataset.</p>
-<p>The classification layer of the pretrained model in <code class="docutils literal notranslate"><span class="pre">state_file</span></code> is
-initilialized from scratch with the classes of the new dataset <code class="docutils literal notranslate"><span class="pre">ds</span></code>.</p>
-<p>The remaining model weights are preserved.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>state_file</strong> (<cite>pathlib.Path</cite>) – Path to a pretrained model.</p></li>
-<li><p><strong>ds</strong> (<cite>pysegcnn.core.dataset.ImageDataset</cite>) – An instance of <cite>pysegcnn.core.dataset.ImageDataset</cite>.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Raises</dt>
-<dd class="field-even"><ul class="simple">
-<li><p><strong>TypeError</strong> – Raised if <code class="docutils literal notranslate"><span class="pre">ds</span></code> is not an instance of
-<cite>pysegcnn.core.dataset.ImageDataset</cite>.</p></li>
-<li><p><strong>ValueError</strong> – Raised if the bands of <code class="docutils literal notranslate"><span class="pre">ds</span></code> do not match the bands of the dataset
-the pretrained model was trained with.</p></li>
-</ul>
-</dd>
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>model</strong> – An instance of <cite>pysegcnn.core.models.Network</cite>. The pretrained model
-adjusted to the new dataset.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>pysegcnn.core.models.Network</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.trainer.NetworkTrainer">
-<em class="property">class </em><code class="sig-prename descclassname">core.trainer.</code><code class="sig-name descname">NetworkTrainer</code><span class="sig-paren">(</span><em class="sig-param">model: pysegcnn.core.models.Network</em>, <em class="sig-param">optimizer: torch.optim.optimizer.Optimizer</em>, <em class="sig-param">loss_function: torch.nn.modules.module.Module</em>, <em class="sig-param">train_dl: torch.utils.data.dataloader.DataLoader</em>, <em class="sig-param">valid_dl: torch.utils.data.dataloader.DataLoader</em>, <em class="sig-param">test_dl: torch.utils.data.dataloader.DataLoader</em>, <em class="sig-param">state_file: pathlib.Path</em>, <em class="sig-param">epochs: int = 1</em>, <em class="sig-param">nthreads: int = 2</em>, <em class="sig-param">early_stop: bool = False</em>, <em class="sig-param">mode: str = 'max'</em>, <em class="sig-param">delta: float = 0</em>, <em class="sig-param">patience: int = 10</em>, <em class="sig-param">checkpoint_state: dict = <factory></em>, <em class="sig-param">save: bool = True</em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/trainer.html#NetworkTrainer"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.NetworkTrainer" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.trainer.BaseConfig" title="core.trainer.BaseConfig"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.trainer.BaseConfig</span></code></a></p>
-<p>Model training class.</p>
-<p>Generic class to train an instance of <cite>pysegcnn.core.models.Network</cite> on
-a dataset of type <cite>pysegcnn.core.dataset.ImageDataset</cite>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>model</strong> (<cite>pysegcnn.core.models.Network</cite>) – The model to train. An instance of <cite>pysegcnn.core.models.Network</cite>.</p></li>
-<li><p><strong>optimizer</strong> (<cite>torch.optim.Optimizer</cite>) – The optimizer to update the model weights. An instance of
-<cite>torch.optim.Optimizer</cite>.</p></li>
-<li><p><strong>loss_function</strong> (<cite>torch.nn.Module</cite>) – The loss function to compute the model error. An instance of
-<cite>torch.nn.Module</cite>.</p></li>
-<li><p><strong>train_dl</strong> (<cite>torch.utils.data.DataLoader</cite>) – The training <cite>torch.utils.data.DataLoader</cite> instance.</p></li>
-<li><p><strong>valid_dl</strong> (<cite>torch.utils.data.DataLoader</cite>) – The validation <cite>torch.utils.data.DataLoader</cite> instance.</p></li>
-<li><p><strong>test_dl</strong> (<cite>torch.utils.data.DataLoader</cite>) – The test <cite>torch.utils.data.DataLoader</cite> instance.</p></li>
-<li><p><strong>state_file</strong> (<cite>pathlib.Path</cite>) – Path to save the model state.</p></li>
-<li><p><strong>epochs</strong> (<cite>int</cite>, optional) – The maximum number of epochs to train. The default is 1.</p></li>
-<li><p><strong>nthreads</strong> (<cite>int</cite>, optional) – The number of cpu threads to use during training. The default is
-torch.get_num_threads().</p></li>
-<li><p><strong>early_stop</strong> (<cite>bool</cite>, optional) – Whether to apply <a href="#id29"><span class="problematic" id="id30">`early stopping`_</span></a>. The default is False.</p></li>
-<li><p><strong>mode</strong> (<cite>str</cite>, optional) – The mode of the early stopping. Depends on the metric measuring
-performance. When using model loss as metric, use <code class="docutils literal notranslate"><span class="pre">mode</span></code> = ‘min’,
-however, when using accuracy as metric, use <code class="docutils literal notranslate"><span class="pre">mode</span></code> = ‘max’. For now,
-only <code class="docutils literal notranslate"><span class="pre">mode</span></code> = ‘max’ is supported. Only used if <code class="docutils literal notranslate"><span class="pre">early_stop</span></code> = True.
-The default is ‘max’.</p></li>
-<li><p><strong>delta</strong> (<cite>float</cite>, optional) – Minimum change in early stopping metric to be considered as an
-improvement. Only used if <code class="docutils literal notranslate"><span class="pre">early_stop</span></code> = True. The default is 0.</p></li>
-<li><p><strong>patience</strong> (<cite>int</cite>, optional) – The number of epochs to wait for an improvement in the early stopping
-metric. If the model does not improve over more than <code class="docutils literal notranslate"><span class="pre">patience</span></code>
-epochs, quit training. Only used if <code class="docutils literal notranslate"><span class="pre">early_stop</span></code> = True.
-The default is 10.</p></li>
-<li><p><strong>checkpoint_state</strong> (<cite>dict</cite> [<cite>str</cite>, <cite>numpy.ndarray</cite>], optional) – <p>A model checkpoint for <code class="docutils literal notranslate"><span class="pre">model</span></code>. If specified, <code class="docutils literal notranslate"><span class="pre">checkpoint_state</span></code>
-should be a dictionary with keys:</p>
-<blockquote>
-<div><dl class="simple">
-<dt><code class="docutils literal notranslate"><span class="pre">'ta'</span></code></dt><dd><p>The accuracy on the training set (<cite>numpy.ndarray</cite>).</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'tl'</span></code></dt><dd><p>The loss on the training set (<cite>numpy.ndarray</cite>).</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'va'</span></code></dt><dd><p>The accuracy on the validation set (<cite>numpy.ndarray</cite>).</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'vl'</span></code></dt><dd><p>The loss on the validation set (<cite>numpy.ndarray</cite>).</p>
-</dd>
-</dl>
-</div></blockquote>
-<p>The default is {}.</p>
-</p></li>
-<li><p><strong>save</strong> (<cite>bool</cite>, optional) – Whether to save the model state to <code class="docutils literal notranslate"><span class="pre">state_file</span></code>. The default is True.</p></li>
-<li><p><strong>_early stopping</strong> (<em>.</em>) – <a class="reference external" href="https://en.wikipedia.org/wiki/Early_stopping">https://en.wikipedia.org/wiki/Early_stopping</a></p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.trainer.NetworkTrainer.predict" title="core.trainer.NetworkTrainer.predict"><code class="xref py py-obj docutils literal notranslate"><span class="pre">predict</span></code></a>()</p></td>
-<td><p>Model inference at training time.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.trainer.NetworkTrainer.save_state" title="core.trainer.NetworkTrainer.save_state"><code class="xref py py-obj docutils literal notranslate"><span class="pre">save_state</span></code></a>()</p></td>
-<td><p>Save the model state.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.trainer.NetworkTrainer.train" title="core.trainer.NetworkTrainer.train"><code class="xref py py-obj docutils literal notranslate"><span class="pre">train</span></code></a>()</p></td>
-<td><p>Train the model.</p></td>
-</tr>
-</tbody>
-</table>
-<dl class="py attribute">
-<dt id="core.trainer.NetworkTrainer.checkpoint_state">
-<code class="sig-name descname">checkpoint_state</code><em class="property">: dict</em><a class="headerlink" href="#core.trainer.NetworkTrainer.checkpoint_state" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.NetworkTrainer.delta">
-<code class="sig-name descname">delta</code><em class="property">: float</em><em class="property"> = 0</em><a class="headerlink" href="#core.trainer.NetworkTrainer.delta" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.NetworkTrainer.early_stop">
-<code class="sig-name descname">early_stop</code><em class="property">: bool</em><em class="property"> = False</em><a class="headerlink" href="#core.trainer.NetworkTrainer.early_stop" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.NetworkTrainer.epochs">
-<code class="sig-name descname">epochs</code><em class="property">: int</em><em class="property"> = 1</em><a class="headerlink" href="#core.trainer.NetworkTrainer.epochs" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.NetworkTrainer.loss_function">
-<code class="sig-name descname">loss_function</code><em class="property">: torch.nn.modules.module.Module</em><a class="headerlink" href="#core.trainer.NetworkTrainer.loss_function" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.NetworkTrainer.mode">
-<code class="sig-name descname">mode</code><em class="property">: str</em><em class="property"> = 'max'</em><a class="headerlink" href="#core.trainer.NetworkTrainer.mode" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.NetworkTrainer.model">
-<code class="sig-name descname">model</code><em class="property">: pysegcnn.core.models.Network</em><a class="headerlink" href="#core.trainer.NetworkTrainer.model" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.NetworkTrainer.nthreads">
-<code class="sig-name descname">nthreads</code><em class="property">: int</em><em class="property"> = 2</em><a class="headerlink" href="#core.trainer.NetworkTrainer.nthreads" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.NetworkTrainer.optimizer">
-<code class="sig-name descname">optimizer</code><em class="property">: torch.optim.optimizer.Optimizer</em><a class="headerlink" href="#core.trainer.NetworkTrainer.optimizer" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.NetworkTrainer.patience">
-<code class="sig-name descname">patience</code><em class="property">: int</em><em class="property"> = 10</em><a class="headerlink" href="#core.trainer.NetworkTrainer.patience" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py method">
-<dt id="core.trainer.NetworkTrainer.predict">
-<code class="sig-name descname">predict</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/trainer.html#NetworkTrainer.predict"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.NetworkTrainer.predict" title="Permalink to this definition">¶</a></dt>
-<dd><p>Model inference at training time.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><ul class="simple">
-<li><p><strong>accuracies</strong> (<cite>numpy.ndarray</cite>) – The mean model prediction accuracy on each mini-batch in the
-validation set.</p></li>
-<li><p><strong>losses</strong> (<cite>numpy.ndarray</cite>) – The model loss for each mini-batch in the validation set.</p></li>
-</ul>
-</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.NetworkTrainer.save">
-<code class="sig-name descname">save</code><em class="property">: bool</em><em class="property"> = True</em><a class="headerlink" href="#core.trainer.NetworkTrainer.save" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py method">
-<dt id="core.trainer.NetworkTrainer.save_state">
-<code class="sig-name descname">save_state</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/trainer.html#NetworkTrainer.save_state"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.NetworkTrainer.save_state" title="Permalink to this definition">¶</a></dt>
-<dd><p>Save the model state.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p></p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p>None.</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.NetworkTrainer.state_file">
-<code class="sig-name descname">state_file</code><em class="property">: pathlib.Path</em><a class="headerlink" href="#core.trainer.NetworkTrainer.state_file" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.NetworkTrainer.test_dl">
-<code class="sig-name descname">test_dl</code><em class="property">: torch.utils.data.dataloader.DataLoader</em><a class="headerlink" href="#core.trainer.NetworkTrainer.test_dl" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py method">
-<dt id="core.trainer.NetworkTrainer.train">
-<code class="sig-name descname">train</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/trainer.html#NetworkTrainer.train"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.NetworkTrainer.train" title="Permalink to this definition">¶</a></dt>
-<dd><p>Train the model.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><p><strong>training_state</strong> – The training state dictionary with keys:
-<code class="docutils literal notranslate"><span class="pre">'ta'</span></code></p>
-<blockquote>
-<div><p>The accuracy on the training set (<cite>numpy.ndarray</cite>).</p>
-</div></blockquote>
-<dl class="simple">
-<dt><code class="docutils literal notranslate"><span class="pre">'tl'</span></code></dt><dd><p>The loss on the training set (<cite>numpy.ndarray</cite>).</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'va'</span></code></dt><dd><p>The accuracy on the validation set (<cite>numpy.ndarray</cite>).</p>
-</dd>
-<dt><code class="docutils literal notranslate"><span class="pre">'vl'</span></code></dt><dd><p>The loss on the validation set (<cite>numpy.ndarray</cite>).</p>
-</dd>
-</dl>
-</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>dict</cite> [<cite>str</cite>, <cite>numpy.ndarray</cite>]</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.NetworkTrainer.train_dl">
-<code class="sig-name descname">train_dl</code><em class="property">: torch.utils.data.dataloader.DataLoader</em><a class="headerlink" href="#core.trainer.NetworkTrainer.train_dl" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.NetworkTrainer.valid_dl">
-<code class="sig-name descname">valid_dl</code><em class="property">: torch.utils.data.dataloader.DataLoader</em><a class="headerlink" href="#core.trainer.NetworkTrainer.valid_dl" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.trainer.SplitConfig">
-<em class="property">class </em><code class="sig-prename descclassname">core.trainer.</code><code class="sig-name descname">SplitConfig</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">split_mode</span><span class="p">:</span> <span class="n">str</span></em>, <em class="sig-param"><span class="n">ttratio</span><span class="p">:</span> <span class="n">float</span></em>, <em class="sig-param"><span class="n">tvratio</span><span class="p">:</span> <span class="n">float</span></em>, <em class="sig-param"><span class="n">date</span><span class="p">:</span> <span class="n">str</span> <span class="o">=</span> <span class="default_value">'yyyymmdd'</span></em>, <em class="sig-param"><span class="n">dateformat</span><span class="p">:</span> <span class="n">str</span> <span class="o">=</span> <span class="default_value">'%Y%m%d'</span></em>, <em class="sig-param"><span class="n">drop</span><span class="p">:</span> <span class="n">float</span> <span class="o">=</span> <span class="default_value">0</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/trainer.html#SplitConfig"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.SplitConfig" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.trainer.BaseConfig" title="core.trainer.BaseConfig"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.trainer.BaseConfig</span></code></a></p>
-<p>Dataset split configuration class.</p>
-<p>Split a dataset into training, validation and test set.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>split_mode</strong> (<cite>str</cite>) – The mode to split the dataset.</p></li>
-<li><p><strong>ttratio</strong> (<cite>float</cite>) – The ratio of training and validation data to test data, e.g.
-<code class="docutils literal notranslate"><span class="pre">ttratio</span></code> = 0.6 means 60% for training and validation, 40% for
-testing.</p></li>
-<li><p><strong>tvratio</strong> (<cite>float</cite>) – The ratio of training data to validation data, e.g. <code class="docutils literal notranslate"><span class="pre">tvratio</span></code> = 0.8
-means 80% training, 20% validation.</p></li>
-<li><p><strong>date</strong> (<cite>str</cite>, optional) – A date. Used if <code class="docutils literal notranslate"><span class="pre">split_mode</span></code> = ‘date’. The default is ‘yyyymmdd’.</p></li>
-<li><p><strong>dateformat</strong> (<cite>str</cite>, optional) – The format of <code class="docutils literal notranslate"><span class="pre">date</span></code>. <code class="docutils literal notranslate"><span class="pre">dateformat</span></code> is used by
-<cite>datetime.datetime.strptime’ to parse ``date`</cite> to a <cite>datetime.datetime</cite>
-object. The default is ‘%Y%m%d’.</p></li>
-<li><p><strong>drop</strong> (<cite>float</cite>, optional) – Whether to drop samples (during training only) with a fraction of
-pixels equal to the constant padding value >= <code class="docutils literal notranslate"><span class="pre">drop</span></code>. <code class="docutils literal notranslate"><span class="pre">drop</span></code> = 0
-means, do not drop any samples. The default is 0.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.trainer.SplitConfig.dataloaders" title="core.trainer.SplitConfig.dataloaders"><code class="xref py py-obj docutils literal notranslate"><span class="pre">dataloaders</span></code></a>(*args, **kwargs)</p></td>
-<td><p>Build <cite>torch.utils.data.DataLoader</cite> instances.</p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="#core.trainer.SplitConfig.train_val_test_split" title="core.trainer.SplitConfig.train_val_test_split"><code class="xref py py-obj docutils literal notranslate"><span class="pre">train_val_test_split</span></code></a>(ds)</p></td>
-<td><p>Split <code class="docutils literal notranslate"><span class="pre">ds</span></code> into training, validation and test set.</p></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="core.trainer.SplitConfig.dataloaders">
-<em class="property">static </em><code class="sig-name descname">dataloaders</code><span class="sig-paren">(</span><em class="sig-param"><span class="o">*</span><span class="n">args</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/trainer.html#SplitConfig.dataloaders"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.SplitConfig.dataloaders" title="Permalink to this definition">¶</a></dt>
-<dd><p>Build <cite>torch.utils.data.DataLoader</cite> instances.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>*args</strong> (<cite>list</cite> [<cite>torch.utils.data.Dataset</cite>]) – List of instances of <cite>torch.utils.data.Dataset</cite>.</p></li>
-<li><p><strong>**kwargs</strong> – Additional keyword arguments passed to
-<cite>torch.utils.data.DataLoader</cite>.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Raises</dt>
-<dd class="field-even"><p><strong>TypeError</strong> – Raised if not each item in <code class="docutils literal notranslate"><span class="pre">args</span></code> is an instance of
-<cite>torch.utils.data.Dataset</cite>.</p>
-</dd>
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>loaders</strong> – List of instances of <cite>torch.utils.data.DataLoader</cite>. If an instance
-of <cite>torch.utils.data.Dataset</cite> in <code class="docutils literal notranslate"><span class="pre">args</span></code> is empty, <cite>None</cite> is
-appended to <code class="docutils literal notranslate"><span class="pre">loaders</span></code> instead of an instance of
-<cite>torch.utils.data.DataLoader</cite>.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>list</cite> [<cite>torch.utils.data.DataLoader</cite>]</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.SplitConfig.date">
-<code class="sig-name descname">date</code><em class="property">: str</em><em class="property"> = 'yyyymmdd'</em><a class="headerlink" href="#core.trainer.SplitConfig.date" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.SplitConfig.dateformat">
-<code class="sig-name descname">dateformat</code><em class="property">: str</em><em class="property"> = '%Y%m%d'</em><a class="headerlink" href="#core.trainer.SplitConfig.dateformat" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.SplitConfig.drop">
-<code class="sig-name descname">drop</code><em class="property">: float</em><em class="property"> = 0</em><a class="headerlink" href="#core.trainer.SplitConfig.drop" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.SplitConfig.split_mode">
-<code class="sig-name descname">split_mode</code><em class="property">: str</em><a class="headerlink" href="#core.trainer.SplitConfig.split_mode" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py method">
-<dt id="core.trainer.SplitConfig.train_val_test_split">
-<code class="sig-name descname">train_val_test_split</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">ds</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/trainer.html#SplitConfig.train_val_test_split"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.SplitConfig.train_val_test_split" title="Permalink to this definition">¶</a></dt>
-<dd><p>Split <code class="docutils literal notranslate"><span class="pre">ds</span></code> into training, validation and test set.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>ds</strong> (<cite>pysegcnn.core.dataset.ImageDataset</cite>) – An instance of <cite>pysegcnn.core.dataset.ImageDataset</cite>.</p>
-</dd>
-<dt class="field-even">Raises</dt>
-<dd class="field-even"><p><strong>TypeError</strong> – Raised if <code class="docutils literal notranslate"><span class="pre">ds</span></code> is not an instance of
-<cite>pysegcnn.core.dataset.ImageDataset</cite>.</p>
-</dd>
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><ul class="simple">
-<li><p><strong>train_ds</strong> (<cite>pysegcnn.core.split.RandomSubset</cite> or)</p></li>
-<li><p><cite>pysegcnn.core.split.SceneSubset</cite>. – The training set.</p></li>
-<li><p><strong>valid_ds</strong> (<cite>pysegcnn.core.split.RandomSubset</cite> or)</p></li>
-<li><p><cite>pysegcnn.core.split.SceneSubset</cite>. – The validation set.</p></li>
-<li><p><strong>test_ds</strong> (<cite>pysegcnn.core.split.RandomSubset</cite> or)</p></li>
-<li><p><cite>pysegcnn.core.split.SceneSubset</cite>. – The test set.</p></li>
-</ul>
-</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.SplitConfig.ttratio">
-<code class="sig-name descname">ttratio</code><em class="property">: float</em><a class="headerlink" href="#core.trainer.SplitConfig.ttratio" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.SplitConfig.tvratio">
-<code class="sig-name descname">tvratio</code><em class="property">: float</em><a class="headerlink" href="#core.trainer.SplitConfig.tvratio" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.trainer.StateConfig">
-<em class="property">class </em><code class="sig-prename descclassname">core.trainer.</code><code class="sig-name descname">StateConfig</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">ds</span><span class="p">:</span> <span class="n">pysegcnn.core.dataset.ImageDataset</span></em>, <em class="sig-param"><span class="n">sc</span><span class="p">:</span> <span class="n"><a class="reference internal" href="#core.trainer.SplitConfig" title="core.trainer.SplitConfig">core.trainer.SplitConfig</a></span></em>, <em class="sig-param"><span class="n">mc</span><span class="p">:</span> <span class="n"><a class="reference internal" href="#core.trainer.ModelConfig" title="core.trainer.ModelConfig">core.trainer.ModelConfig</a></span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/trainer.html#StateConfig"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.StateConfig" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.trainer.BaseConfig" title="core.trainer.BaseConfig"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.trainer.BaseConfig</span></code></a></p>
-<p>Model state configuration class.</p>
-<p>Generate the model state filename according to the following naming
-convention:</p>
-<p>model_dataset_optimizer_splitmode_splitparams_tilesize_batchsize_bands.pt</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>ds</strong> (<cite>pysegcnn.core.dataset.ImageDataset</cite>) – An instance of <cite>pysegcnn.core.dataset.ImageDataset</cite>.</p></li>
-<li><p><strong>sc</strong> (<cite>pysegcnn.core.trainer.SplitConfig</cite>) – An instance of <cite>pysegcnn.core.trainer.SplitConfig</cite>.</p></li>
-<li><p><strong>mc</strong> (<cite>pysegcnn.core.trainer.ModelConfig</cite>) – An instance of <cite>pysegcnn.core.trainer.SplitConfig</cite>.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><a class="reference internal" href="#core.trainer.StateConfig.init_state" title="core.trainer.StateConfig.init_state"><code class="xref py py-obj docutils literal notranslate"><span class="pre">init_state</span></code></a>()</p></td>
-<td><p>Generate the model state filename.</p></td>
-</tr>
-</tbody>
-</table>
-<dl class="py attribute">
-<dt id="core.trainer.StateConfig.ds">
-<code class="sig-name descname">ds</code><em class="property">: pysegcnn.core.dataset.ImageDataset</em><a class="headerlink" href="#core.trainer.StateConfig.ds" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py method">
-<dt id="core.trainer.StateConfig.init_state">
-<code class="sig-name descname">init_state</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/trainer.html#StateConfig.init_state"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.trainer.StateConfig.init_state" title="Permalink to this definition">¶</a></dt>
-<dd><p>Generate the model state filename.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>state</strong> – The path to the model state file.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>pathlib.Path</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.StateConfig.mc">
-<code class="sig-name descname">mc</code><em class="property">: <a class="reference internal" href="#core.trainer.ModelConfig" title="core.trainer.ModelConfig">core.trainer.ModelConfig</a></em><a class="headerlink" href="#core.trainer.StateConfig.mc" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-<dl class="py attribute">
-<dt id="core.trainer.StateConfig.sc">
-<code class="sig-name descname">sc</code><em class="property">: <a class="reference internal" href="#core.trainer.SplitConfig" title="core.trainer.SplitConfig">core.trainer.SplitConfig</a></em><a class="headerlink" href="#core.trainer.StateConfig.sc" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-</dd></dl>
-
-</div>
-</div>
-<div class="section" id="module-core.transforms">
-<span id="core-transforms-module"></span><h2>core.transforms module<a class="headerlink" href="#module-core.transforms" title="Permalink to this headline">¶</a></h2>
-<p>Data augmentation.</p>
-<p>This module provides classes implementing common image augmentation methods.</p>
-<p>These methods may be used to artificially increase a dataset.</p>
-<div class="section" id="id11">
-<h3>License<a class="headerlink" href="#id11" title="Permalink to this headline">¶</a></h3>
-<blockquote>
-<div><p>Copyright (c) 2020 Daniel Frisinghelli</p>
-<p>This source code is licensed under the GNU General Public License v3.</p>
-<p>See the LICENSE file in the repository’s root directory.</p>
-</div></blockquote>
-<dl class="py class">
-<dt id="core.transforms.Augment">
-<em class="property">class </em><code class="sig-prename descclassname">core.transforms.</code><code class="sig-name descname">Augment</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">transforms</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/transforms.html#Augment"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.transforms.Augment" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p>
-<p>Apply a sequence of transformations.</p>
-<p>Container class applying each transformation in <code class="docutils literal notranslate"><span class="pre">transforms</span></code> in order.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>transforms</strong> (<cite>list</cite> or <cite>tuple</cite>) – A sequence of instances of <cite>pysegcnn.core.transforms.VariantTransform</cite>
-or <cite>pysegcnn.core.transforms.InvariantTransform</cite>.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">__call__</span></code>(image, gt)</p></td>
-<td><p>Apply a sequence of transformations to <code class="docutils literal notranslate"><span class="pre">image</span></code>.</p></td>
-</tr>
-</tbody>
-</table>
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.transforms.FlipLr">
-<em class="property">class </em><code class="sig-prename descclassname">core.transforms.</code><code class="sig-name descname">FlipLr</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">p</span><span class="o">=</span><span class="default_value">0.5</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/transforms.html#FlipLr"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.transforms.FlipLr" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.transforms.VariantTransform" title="core.transforms.VariantTransform"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.transforms.VariantTransform</span></code></a></p>
-<p>Flip an image horizontally.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>p</strong> (<cite>float</cite>, optional) – The probability to apply the transformation. The default is 0.5.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">__call__</span></code>(image)</p></td>
-<td><p>Apply transformation.</p></td>
-</tr>
-</tbody>
-</table>
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.transforms.FlipUd">
-<em class="property">class </em><code class="sig-prename descclassname">core.transforms.</code><code class="sig-name descname">FlipUd</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">p</span><span class="o">=</span><span class="default_value">0.5</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/transforms.html#FlipUd"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.transforms.FlipUd" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.transforms.VariantTransform" title="core.transforms.VariantTransform"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.transforms.VariantTransform</span></code></a></p>
-<p>Flip an image vertically.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>p</strong> (<cite>float</cite>, optional) – The probability to apply the transformation. The default is 0.5.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">__call__</span></code>(image)</p></td>
-<td><p>Apply transformation.</p></td>
-</tr>
-</tbody>
-</table>
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.transforms.InvariantTransform">
-<em class="property">class </em><code class="sig-prename descclassname">core.transforms.</code><code class="sig-name descname">InvariantTransform</code><a class="reference internal" href="../_modules/core/transforms.html#InvariantTransform"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.transforms.InvariantTransform" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.transforms.Transform" title="core.transforms.Transform"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.transforms.Transform</span></code></a></p>
-<p>Base class for a spatially invariant transformation.</p>
-<p>Transformation on the ground truth not required.</p>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">__call__</span></code>(image)</p></td>
-<td><p>Apply transformation.</p></td>
-</tr>
-</tbody>
-</table>
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.transforms.Noise">
-<em class="property">class </em><code class="sig-prename descclassname">core.transforms.</code><code class="sig-name descname">Noise</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">mode</span></em>, <em class="sig-param"><span class="n">mean</span><span class="o">=</span><span class="default_value">0</span></em>, <em class="sig-param"><span class="n">var</span><span class="o">=</span><span class="default_value">0.05</span></em>, <em class="sig-param"><span class="n">p</span><span class="o">=</span><span class="default_value">0.5</span></em>, <em class="sig-param"><span class="n">exclude</span><span class="o">=</span><span class="default_value">[]</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/transforms.html#Noise"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.transforms.Noise" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.transforms.InvariantTransform" title="core.transforms.InvariantTransform"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.transforms.InvariantTransform</span></code></a></p>
-<p>Add gaussian noise to an image.</p>
-<p>Valid modes are:</p>
-<blockquote>
-<div><p>‘add’: image = image + noise
-‘speckle’ : image = image + image * noise</p>
-</div></blockquote>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>mode</strong> (<cite>str</cite>) – The mode to add the noise.</p></li>
-<li><p><strong>mean</strong> (<cite>float</cite>, optional) – The mean of the gaussian distribution from which the noise is sampled.
-The default is 0.</p></li>
-<li><p><strong>var</strong> (<cite>float</cite>, optional) – The variance of the gaussian distribution from which the noise is
-sampled. The default is 0.05.</p></li>
-<li><p><strong>p</strong> (<cite>float</cite>, optional) – The probability to apply the transformation. The default is 0.5.</p></li>
-<li><p><strong>exclude</strong> (<cite>list</cite> [<cite>float</cite>] or <cite>list</cite> [<cite>int</cite>], optional) – Values for which the noise is not added. Useful for pixels resulting
-from image padding. The default is [].</p></li>
-</ul>
-</dd>
-<dt class="field-even">Raises</dt>
-<dd class="field-even"><p><strong>ValueError</strong> – Raised if <code class="docutils literal notranslate"><span class="pre">mode</span></code> is not supported.</p>
-</dd>
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p></p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">__call__</span></code>(image)</p></td>
-<td><p>Apply transformation.</p></td>
-</tr>
-</tbody>
-</table>
-<dl class="py attribute">
-<dt id="core.transforms.Noise.modes">
-<code class="sig-name descname">modes</code><em class="property"> = ['add', 'speckle']</em><a class="headerlink" href="#core.transforms.Noise.modes" title="Permalink to this definition">¶</a></dt>
-<dd></dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.transforms.Rotate">
-<em class="property">class </em><code class="sig-prename descclassname">core.transforms.</code><code class="sig-name descname">Rotate</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">angle</span></em>, <em class="sig-param"><span class="n">p</span><span class="o">=</span><span class="default_value">0.5</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/transforms.html#Rotate"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.transforms.Rotate" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.transforms.VariantTransform" title="core.transforms.VariantTransform"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.transforms.VariantTransform</span></code></a></p>
-<p>Rotate an image by <code class="docutils literal notranslate"><span class="pre">angle</span></code>.</p>
-<p>The image is rotated in the spatial plane.</p>
-<p>If the input array has more then two dimensions, the spatial dimensions are
-assumed to be the last two dimensions of the array.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>angle</strong> (<cite>float</cite>) – The rotation angle in degrees.</p></li>
-<li><p><strong>p</strong> (<cite>float</cite>, optional) – The probability to apply the transformation. The default is 0.5.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">__call__</span></code>(image)</p></td>
-<td><p>Apply transformation.</p></td>
-</tr>
-</tbody>
-</table>
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.transforms.Transform">
-<em class="property">class </em><code class="sig-prename descclassname">core.transforms.</code><code class="sig-name descname">Transform</code><a class="reference internal" href="../_modules/core/transforms.html#Transform"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.transforms.Transform" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p>
-<p>Base class for an image transformation.</p>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">__call__</span></code>(image)</p></td>
-<td><p>Apply transformation.</p></td>
-</tr>
-</tbody>
-</table>
-</dd></dl>
-
-<dl class="py class">
-<dt id="core.transforms.VariantTransform">
-<em class="property">class </em><code class="sig-prename descclassname">core.transforms.</code><code class="sig-name descname">VariantTransform</code><a class="reference internal" href="../_modules/core/transforms.html#VariantTransform"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.transforms.VariantTransform" title="Permalink to this definition">¶</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#core.transforms.Transform" title="core.transforms.Transform"><code class="xref py py-class docutils literal notranslate"><span class="pre">core.transforms.Transform</span></code></a></p>
-<p>Base class for a spatially variant transformation.</p>
-<p>Transformation on the ground truth required.</p>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">__call__</span></code>(image)</p></td>
-<td><p>Apply transformation.</p></td>
-</tr>
-</tbody>
-</table>
-</dd></dl>
-
-</div>
-</div>
-<div class="section" id="module-core.utils">
-<span id="core-utils-module"></span><h2>core.utils module<a class="headerlink" href="#module-core.utils" title="Permalink to this headline">¶</a></h2>
-<p>Utility functions mainly for image IO and reshaping.</p>
-<div class="section" id="id12">
-<h3>License<a class="headerlink" href="#id12" title="Permalink to this headline">¶</a></h3>
-<blockquote>
-<div><p>Copyright (c) 2020 Daniel Frisinghelli</p>
-<p>This source code is licensed under the GNU General Public License v3.</p>
-<p>See the LICENSE file in the repository’s root directory.</p>
-</div></blockquote>
-<dl class="py function">
-<dt id="core.utils.accuracy_function">
-<code class="sig-prename descclassname">core.utils.</code><code class="sig-name descname">accuracy_function</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">outputs</span></em>, <em class="sig-param"><span class="n">labels</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/utils.html#accuracy_function"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.utils.accuracy_function" title="Permalink to this definition">¶</a></dt>
-<dd><p>Calculate prediction accuracy.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>outputs</strong> (<cite>torch.Tensor</cite> or array_like) – The model prediction.</p></li>
-<li><p><strong>labels</strong> (<cite>torch.Tensor</cite> or array_like) – The ground truth.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>accuracy</strong> – Mean prediction accuracy.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>float</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt id="core.utils.check_tile_extend">
-<code class="sig-prename descclassname">core.utils.</code><code class="sig-name descname">check_tile_extend</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">img_size</span></em>, <em class="sig-param"><span class="n">topleft</span></em>, <em class="sig-param"><span class="n">tile_size</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/utils.html#check_tile_extend"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.utils.check_tile_extend" title="Permalink to this definition">¶</a></dt>
-<dd><p>Check if a tile exceeds the image size.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>img_size</strong> (<cite>tuple</cite>) – The image size (height, width).</p></li>
-<li><p><strong>topleft</strong> (<cite>tuple</cite>) – The topleft corner of the tile (y, x).</p></li>
-<li><p><strong>tile_size</strong> (<cite>int</cite>) – The size of the tile.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><ul class="simple">
-<li><p><strong>nrows</strong> (<cite>int</cite>) – Number of rows of the tile within the image.</p></li>
-<li><p><strong>ncols</strong> (<em>TYPE</em>) – Number of columns of the tile within the image.</p></li>
-</ul>
-</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt id="core.utils.destack_tiff">
-<code class="sig-prename descclassname">core.utils.</code><code class="sig-name descname">destack_tiff</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">image</span></em>, <em class="sig-param"><span class="n">outpath</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="n">overwrite</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">remove</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">suffix</span><span class="o">=</span><span class="default_value">''</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/utils.html#destack_tiff"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.utils.destack_tiff" title="Permalink to this definition">¶</a></dt>
-<dd><p>Destack a TIFF with more than one band into a TIFF file for each band.</p>
-<p>Each band in <code class="docutils literal notranslate"><span class="pre">image</span></code> is saved to <code class="docutils literal notranslate"><span class="pre">outpath</span></code> as distinct TIFF file.
-The default filenames are: “filename(<code class="docutils literal notranslate"><span class="pre">image</span></code>) + _B(i).tifâ€, where i is
-the respective number of each band in <code class="docutils literal notranslate"><span class="pre">image</span></code>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>image</strong> (<cite>str</cite> or <cite>pathlib.Path</cite>) – The TIFF to destack.</p></li>
-<li><p><strong>outpath</strong> (<cite>str</cite>, optional) – Path to save the output TIFF files. The default is None. If None,
-<code class="docutils literal notranslate"><span class="pre">outpath</span></code> is the path to <code class="docutils literal notranslate"><span class="pre">image</span></code>.</p></li>
-<li><p><strong>remove</strong> (<cite>bool</cite>, optional) – Whether to remove <code class="docutils literal notranslate"><span class="pre">image</span></code> from disk after destacking. The default is
-False.</p></li>
-<li><p><strong>overwrite</strong> (<cite>bool</cite>, optional) – Whether to overwrite existing TIFF files.</p></li>
-<li><p><strong>suffix</strong> (<cite>str</cite>, optional) – String to append to the filename of <code class="docutils literal notranslate"><span class="pre">image</span></code>. If specified, the TIFF
-filenames for each band in <code class="docutils literal notranslate"><span class="pre">image</span></code> are, “filename(<code class="docutils literal notranslate"><span class="pre">image</span></code>) +
-+ _B(i)_ + <code class="docutils literal notranslate"><span class="pre">suffix</span></code>.tifâ€. The default is ‘’.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Raises</dt>
-<dd class="field-even"><p><strong>FileNotFoundError</strong> – Raised if <code class="docutils literal notranslate"><span class="pre">image</span></code> does not exist.</p>
-</dd>
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p></p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p>None.</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt id="core.utils.doy2date">
-<code class="sig-prename descclassname">core.utils.</code><code class="sig-name descname">doy2date</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">year</span></em>, <em class="sig-param"><span class="n">doy</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/utils.html#doy2date"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.utils.doy2date" title="Permalink to this definition">¶</a></dt>
-<dd><p>Convert the (year, day of the year) date format to a datetime object.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>year</strong> (<cite>int</cite>) – The year</p></li>
-<li><p><strong>doy</strong> (<cite>int</cite>) – The day of the year</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>date</strong> – The converted date.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>datetime.datetime</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt id="core.utils.extract_archive">
-<code class="sig-prename descclassname">core.utils.</code><code class="sig-name descname">extract_archive</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">inpath</span></em>, <em class="sig-param"><span class="n">outpath</span></em>, <em class="sig-param"><span class="n">overwrite</span><span class="o">=</span><span class="default_value">False</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/utils.html#extract_archive"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.utils.extract_archive" title="Permalink to this definition">¶</a></dt>
-<dd><p>Extract files from an archive.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>inpath</strong> (<cite>str</cite> or <cite>pathlib.Path</cite>) – Path to an archive.</p></li>
-<li><p><strong>outpath</strong> (<cite>str</cite> or <cite>pathlib.Path</cite>) – Path to save extracted files.</p></li>
-<li><p><strong>overwrite</strong> (<cite>bool</cite>, optional) – Whether to overwrite existing extracted files.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>subdir</strong> – path to the extracted files</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>str</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt id="core.utils.get_radiometric_constants">
-<code class="sig-prename descclassname">core.utils.</code><code class="sig-name descname">get_radiometric_constants</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">metadata</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/utils.html#get_radiometric_constants"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.utils.get_radiometric_constants" title="Permalink to this definition">¶</a></dt>
-<dd><p>Retrieve the radiometric calibration constants.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>metadata</strong> (<cite>dict</cite>) – The dictionary returned by <code class="docutils literal notranslate"><span class="pre">read_landsat_metadata</span></code>.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><ul class="simple">
-<li><p><strong>oli</strong> (<cite>dict</cite>) – Radiometric rescaling factors of the OLI sensor.</p></li>
-<li><p><strong>tir</strong> (<cite>dict</cite>) – Thermal conversion constants of the TIRS sensor.</p></li>
-</ul>
-</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt id="core.utils.img2np">
-<code class="sig-prename descclassname">core.utils.</code><code class="sig-name descname">img2np</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">path</span></em>, <em class="sig-param"><span class="n">tile_size</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="n">tile</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="n">pad</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">cval</span><span class="o">=</span><span class="default_value">0</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/utils.html#img2np"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.utils.img2np" title="Permalink to this definition">¶</a></dt>
-<dd><p>Read an image to a <cite>numpy.ndarray</cite>.</p>
-<p>If <code class="docutils literal notranslate"><span class="pre">tile_size</span></code> is not <cite>None</cite>, the input image is divided into square
-tiles of size (<code class="docutils literal notranslate"><span class="pre">tile_size</span></code>, <code class="docutils literal notranslate"><span class="pre">tile_size</span></code>). If the image is not evenly
-divisible and <code class="docutils literal notranslate"><span class="pre">pad</span></code> = False, a <cite>ValueError</cite> is raised. However, if
-<code class="docutils literal notranslate"><span class="pre">pad</span></code> = True, center padding with constant value <code class="docutils literal notranslate"><span class="pre">cval</span></code> is applied.</p>
-<p>The tiling works as follows:</p>
-<blockquote>
-<div><p>(Padded) Input image:</p>
-<div class="line-block">
-<div class="line-block">
-<div class="line">| | | |</div>
-</div>
-<div class="line">tile_00 | tile_01 | … | tile_0n |</div>
-<div class="line-block">
-<div class="line">| | | |</div>
-</div>
-</div>
-<p><a href="#id17"><span class="problematic" id="id18">|----------------------------------------------|</span></a>
-| | | | |
-| tile_10 | tile_11 | … | tile_1n |
-| | | | |
-<a href="#id19"><span class="problematic" id="id20">|----------------------------------------------|</span></a>
-| | | | |
-| … | … | … | … |
-| | | | |
-<a href="#id21"><span class="problematic" id="id22">|----------------------------------------------|</span></a>
-| | | | |
-| tile_m0 | tile_m1 | … | tile_mn |
-| | | | |
-————————————————</p>
-</div></blockquote>
-<p>where m = n. Each tile has its id, which starts at 0 in the topleft corner
-of the input image, i.e. tile_00 has id=0, and increases along the width
-axis, i.e. tile_0n has id=n, tile_10 has id=n+1, …, tile_mn has
-id=(m * n) - 1.</p>
-<p>If <code class="docutils literal notranslate"><span class="pre">tile</span></code> is an integer, only the tile with id = <code class="docutils literal notranslate"><span class="pre">tile</span></code> is returned.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>path</strong> (<cite>str</cite> or <cite>None</cite> or <cite>numpy.ndarray</cite>) – The image to read.</p></li>
-<li><p><strong>tile_size</strong> (<cite>None</cite> or <cite>int</cite>, optional) – The size of a tile. The default is None.</p></li>
-<li><p><strong>tile</strong> (<cite>int</cite>, optional) – The tile id. The default is None.</p></li>
-<li><p><strong>pad</strong> (<cite>bool</cite>, optional) – Whether to center pad the input image. The default is False.</p></li>
-<li><p><strong>cval</strong> (<cite>float</cite>, optional) – The constant padding value. The default is 0.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Raises</dt>
-<dd class="field-even"><ul class="simple">
-<li><p><strong>FileNotFoundError</strong> – Raised if <code class="docutils literal notranslate"><span class="pre">path</span></code> is a path that does not exist.</p></li>
-<li><p><strong>TypeError</strong> – Raised if <code class="docutils literal notranslate"><span class="pre">path</span></code> is not <cite>str</cite> or <cite>None</cite> or <cite>numpy.ndarray</cite>.</p></li>
-</ul>
-</dd>
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><p><strong>image</strong> –</p>
-<p>The image array. The output shape is:</p>
-<blockquote>
-<div><dl>
-<dt>if <code class="docutils literal notranslate"><span class="pre">tile_size</span></code> is not <cite>None</cite>:</dt><dd><p>shape=(tiles, bands, tile_size, tile_size)
-if the image does only have one band:</p>
-<blockquote>
-<div><p>shape=(tiles, tile_size, tile_size)</p>
-</div></blockquote>
-</dd>
-<dt>else:</dt><dd><p>shape=(bands, height, width)
-if the image does only have one band:</p>
-<blockquote>
-<div><p>shape=(height, width)</p>
-</div></blockquote>
-</dd>
-</dl>
-</div></blockquote>
-</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>numpy.ndarray</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt id="core.utils.is_divisible">
-<code class="sig-prename descclassname">core.utils.</code><code class="sig-name descname">is_divisible</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">img_size</span></em>, <em class="sig-param"><span class="n">tile_size</span></em>, <em class="sig-param"><span class="n">pad</span><span class="o">=</span><span class="default_value">False</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/utils.html#is_divisible"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.utils.is_divisible" title="Permalink to this definition">¶</a></dt>
-<dd><p>Check whether an image is evenly divisible into square tiles.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>img_size</strong> (<cite>tuple</cite>) – The image size (height, width).</p></li>
-<li><p><strong>tile_size</strong> (<cite>int</cite>) – The size of the tile.</p></li>
-<li><p><strong>pad</strong> (<cite>bool</cite>, optional) – Whether to center pad the input image. The default is False.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Raises</dt>
-<dd class="field-even"><p><strong>ValueError</strong> – Raised if the image is not evenly divisible and <code class="docutils literal notranslate"><span class="pre">pad</span></code> = False.</p>
-</dd>
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><ul class="simple">
-<li><p><strong>ntiles</strong> (<cite>int</cite>) – The number of tiles fitting <code class="docutils literal notranslate"><span class="pre">img_size</span></code>.</p></li>
-<li><p><strong>padding</strong> (<cite>tuple</cite>) – The amount of padding (bottom, left, top, right).</p></li>
-</ul>
-</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt id="core.utils.item_in_enum">
-<code class="sig-prename descclassname">core.utils.</code><code class="sig-name descname">item_in_enum</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">name</span></em>, <em class="sig-param"><span class="n">enum</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/utils.html#item_in_enum"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.utils.item_in_enum" title="Permalink to this definition">¶</a></dt>
-<dd><p>Check if an item exists in an enumeration.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>name</strong> (<cite>str</cite>) – Name of the item.</p></li>
-<li><p><strong>enum</strong> (<cite>enum.Enum</cite>) – An instance of <cite>enum.Enum</cite>.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Raises</dt>
-<dd class="field-even"><p><strong>ValueError</strong> – Raised if <code class="docutils literal notranslate"><span class="pre">name</span></code> is not in <code class="docutils literal notranslate"><span class="pre">enum</span></code>.</p>
-</dd>
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p>The value of <code class="docutils literal notranslate"><span class="pre">name</span></code> in <code class="docutils literal notranslate"><span class="pre">enum</span></code>.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p>value</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt id="core.utils.landsat_radiometric_calibration">
-<code class="sig-prename descclassname">core.utils.</code><code class="sig-name descname">landsat_radiometric_calibration</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">scene</span></em>, <em class="sig-param"><span class="n">outpath</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="n">exclude</span><span class="o">=</span><span class="default_value">[]</span></em>, <em class="sig-param"><span class="n">radiance</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">overwrite</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">remove_raw</span><span class="o">=</span><span class="default_value">True</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/utils.html#landsat_radiometric_calibration"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.utils.landsat_radiometric_calibration" title="Permalink to this definition">¶</a></dt>
-<dd><p>Radiometric calibration of Landsat Collection Level 1 scenes.</p>
-<p>Convert the Landsat OLI bands to top of atmosphere radiance or reflectance
-and the TIRS bands to top of atmosphere brightness temperature.</p>
-<p>Conversion is performed following the <a href="#id31"><span class="problematic" id="id32">`equations`_</span></a> provided by the USGS.</p>
-<p>The filename of each band is extended by one of the following suffixes,
-depending on the type of the radiometric calibration:</p>
-<blockquote>
-<div><p>‘toa_ref’: top of atmosphere reflectance
-‘toa_rad’: top of atmopshere radiance
-‘toa_brt’: top of atmosphere brightness temperature</p>
-</div></blockquote>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>scene</strong> (<cite>str</cite> or <cite>pathlib.Path</cite>) – Path to a Landsat scene in digital number format.</p></li>
-<li><p><strong>outpath</strong> (<cite>str</cite> or <cite>pathlib.Path</cite>, optional) – Path to save the calibrated images. The default is None, which means
-saving to <code class="docutils literal notranslate"><span class="pre">scene</span></code>.</p></li>
-<li><p><strong>exclude</strong> (<cite>list</cite> [<cite>str</cite>], optional) – Bands to exclude from the radiometric calibration. The default is [].</p></li>
-<li><p><strong>radiance</strong> (<cite>bool</cite>, optional) – Whether to calculate top of atmosphere radiance. The default is False,
-which means calculating top of atmopshere reflectance.</p></li>
-<li><p><strong>overwrite</strong> (<cite>bool</cite>, optional) – Whether to overwrite the calibrated images. The default is False.</p></li>
-<li><p><strong>remove_raw</strong> (<cite>bool</cite>, optional) – Whether to remove the raw digitial number images. The default is True.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Raises</dt>
-<dd class="field-even"><p><strong>FileNotFoundError</strong> – Raised if <code class="docutils literal notranslate"><span class="pre">scene</span></code> does not exist.</p>
-</dd>
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><ul class="simple">
-<li><p><em>None.</em></p></li>
-<li><p><em>.. _equations</em> – <a class="reference external" href="https://www.usgs.gov/land-resources/nli/landsat/using-usgs-landsat-level-1-data-product">https://www.usgs.gov/land-resources/nli/landsat/using-usgs-landsat-level-1-data-product</a></p></li>
-</ul>
-</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt id="core.utils.parse_landsat_scene">
-<code class="sig-prename descclassname">core.utils.</code><code class="sig-name descname">parse_landsat_scene</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">scene_id</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/utils.html#parse_landsat_scene"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.utils.parse_landsat_scene" title="Permalink to this definition">¶</a></dt>
-<dd><p>Parse a Landsat scene identifier.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>scene_id</strong> (<cite>str</cite>) – A Landsat scene identifier.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>scene</strong> – A dictionary containing scene metadata. If <cite>None</cite>, <code class="docutils literal notranslate"><span class="pre">scene_id</span></code> is not
-a valid Landsat scene identifier.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>dict</cite> or <cite>None</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt id="core.utils.parse_sentinel2_scene">
-<code class="sig-prename descclassname">core.utils.</code><code class="sig-name descname">parse_sentinel2_scene</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">scene_id</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/utils.html#parse_sentinel2_scene"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.utils.parse_sentinel2_scene" title="Permalink to this definition">¶</a></dt>
-<dd><p>Parse a Sentinel-2 scene identifier.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>scene_id</strong> (<cite>str</cite>) – A Sentinel-2 scene identifier.</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>scene</strong> – A dictionary containing scene metadata. If <cite>None</cite>, <code class="docutils literal notranslate"><span class="pre">scene_id</span></code> is not
-a valid Sentinel-2 scene identifier.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>dict</cite> or <cite>None</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt id="core.utils.read_landsat_metadata">
-<code class="sig-prename descclassname">core.utils.</code><code class="sig-name descname">read_landsat_metadata</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">file</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/utils.html#read_landsat_metadata"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.utils.read_landsat_metadata" title="Permalink to this definition">¶</a></dt>
-<dd><p>Parse the Landsat metadata <a href="#id13"><span class="problematic" id="id14">*</span></a>_MTL.txt file.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>file</strong> (<cite>str</cite> or <cite>pathlib.Path</cite>) – Path to a Landsat <a href="#id15"><span class="problematic" id="id16">*</span></a>_MTL.txt file.</p>
-</dd>
-<dt class="field-even">Raises</dt>
-<dd class="field-even"><p><strong>FileNotFoundError</strong> – Raised if <code class="docutils literal notranslate"><span class="pre">file</span></code> does not exist.</p>
-</dd>
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p><strong>metadata</strong> – The metadata text file as dictionary, where each line is a (key, value)
-pair.</p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p><cite>dict</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt id="core.utils.reconstruct_scene">
-<code class="sig-prename descclassname">core.utils.</code><code class="sig-name descname">reconstruct_scene</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">tiles</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/utils.html#reconstruct_scene"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.utils.reconstruct_scene" title="Permalink to this definition">¶</a></dt>
-<dd><p>Reconstruct a tiled image.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>tiles</strong> (<cite>torch.Tensor</cite> or <cite>numpy.ndarray</cite>) – The tiled image, shape=(tiles, bands, tile_size, tile_size) or
-shape=(tiles, tile_size, tile_size).</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>image</strong> – The reconstructed image.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>numpy.ndarray</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt id="core.utils.standard_eo_structure">
-<code class="sig-prename descclassname">core.utils.</code><code class="sig-name descname">standard_eo_structure</code><span class="sig-paren">(</span><em class="sig-param">source_path</em>, <em class="sig-param">target_path</em>, <em class="sig-param">overwrite=False</em>, <em class="sig-param">move=False</em>, <em class="sig-param">parser=<function parse_landsat_scene></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/utils.html#standard_eo_structure"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.utils.standard_eo_structure" title="Permalink to this definition">¶</a></dt>
-<dd><p>Modify the directory structure of a remote sensing dataset.</p>
-<p>This function assumes that <code class="docutils literal notranslate"><span class="pre">source_path</span></code> points to a directory containing
-remote sensing data, where each file in <code class="docutils literal notranslate"><span class="pre">source_path</span></code> and its sub-folders
-should contain a scene identifier in its filename. The scene identifier is
-used to restructure the dataset.</p>
-<p>Currently, Landsat and Sentinel-2 datasets are supported.</p>
-<p>The directory tree in <code class="docutils literal notranslate"><span class="pre">source_path</span></code> is modified to the following
-structure in <code class="docutils literal notranslate"><span class="pre">target_path</span></code>:</p>
-<blockquote>
-<div><dl>
-<dt>target_path/</dt><dd><dl class="simple">
-<dt>scene_id_1/</dt><dd><p>files matching scene_id_1</p>
-</dd>
-<dt>scene_id_2/</dt><dd><p>files matching scene_id_2</p>
-</dd>
-</dl>
-<p>.
-scene_id_n/</p>
-<blockquote>
-<div><p>files matching scene_id_n</p>
-</div></blockquote>
-</dd>
-</dl>
-</div></blockquote>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>source_path</strong> (<cite>str</cite> or <cite>pathlib.Path</cite>) – Path to the remote sensing dataset.</p></li>
-<li><p><strong>target_path</strong> (<cite>str</cite> or <cite>pathlib.Path</cite>) – Path to save the restructured dataset.</p></li>
-<li><p><strong>overwrite</strong> (<cite>bool</cite>, optional) – Whether to overwrite existing files in <code class="docutils literal notranslate"><span class="pre">target_path</span></code>.
-The default is True.</p></li>
-<li><p><strong>move</strong> (<cite>bool</cite>, optional) – Whether to move the files from <code class="docutils literal notranslate"><span class="pre">source_path</span></code> to <code class="docutils literal notranslate"><span class="pre">target_path</span></code>. If
-True, files in <code class="docutils literal notranslate"><span class="pre">source_path</span></code> are moved to <code class="docutils literal notranslate"><span class="pre">target_path</span></code>, if False,
-files in <code class="docutils literal notranslate"><span class="pre">source_path</span></code> are copied to <code class="docutils literal notranslate"><span class="pre">target_path</span></code>. The default is
-False.</p></li>
-<li><p><strong>parser</strong> (<cite>function</cite>, optional) – The scene identifier parsing function. Depends on the sensor of the
-dataset. See e.g., <cite>pysegcnn.core.utils.parse_landsat_scene</cite>.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt id="core.utils.tile_topleft_corner">
-<code class="sig-prename descclassname">core.utils.</code><code class="sig-name descname">tile_topleft_corner</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">img_size</span></em>, <em class="sig-param"><span class="n">tile_size</span></em><span class="sig-paren">)</span><a class="reference internal" href="../_modules/core/utils.html#tile_topleft_corner"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#core.utils.tile_topleft_corner" title="Permalink to this definition">¶</a></dt>
-<dd><p>Return the topleft corners of the tiles in the image.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>img_size</strong> (<cite>tuple</cite>) – The image size (height, width).</p></li>
-<li><p><strong>tile_size</strong> (<cite>int</cite>) – The size of the tile.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p><strong>indices</strong> – The keys of <code class="docutils literal notranslate"><span class="pre">indices</span></code> are the tile ids (<cite>int</cite>) and the values are the
-topleft corners (<cite>tuple</cite> = (y, x)) of the tiles.</p>
-</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p><cite>dict</cite></p>
-</dd>
-</dl>
-</dd></dl>
-
-</div>
-</div>
-<div class="section" id="module-core">
-<span id="module-contents"></span><h2>Module contents<a class="headerlink" href="#module-core" title="Permalink to this headline">¶</a></h2>
-</div>
-</div>
-
-
- </div>
-
- </div>
- <footer>
-
- <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
-
- <a href="main.html" class="btn btn-neutral float-right" title="main package" accesskey="n" rel="next">Next <span class="fa fa-arrow-circle-right"></span></a>
-
-
- <a href="modules.html" class="btn btn-neutral float-left" title="pysegcnn" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left"></span> Previous</a>
-
- </div>
-
-
- <hr/>
-
- <div role="contentinfo">
- <p>
-
- © Copyright 2020, Daniel Frisinghelli
-
- </p>
- </div>
-
-
-
- Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a
-
- <a href="https://github.com/rtfd/sphinx_rtd_theme">theme</a>
-
- provided by <a href="https://readthedocs.org">Read the Docs</a>.
-
-</footer>
-
- </div>
- </div>
-
- </section>
-
- </div>
-
-
- <script type="text/javascript">
- jQuery(function () {
- SphinxRtdTheme.Navigation.enable(true);
- });
- </script>
-
-
-
-
-
-
-</body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/source/generated/pysegcnn.core.dataset.Cloud95Dataset.html b/docs/_build/html/source/generated/pysegcnn.core.dataset.Cloud95Dataset.html
index 71193ba899ffa348b9a91465bd75b0563c14462b..78fe3ebf4be34eb40a674e610a52ff4928fa76dc 100644
--- a/docs/_build/html/source/generated/pysegcnn.core.dataset.Cloud95Dataset.html
+++ b/docs/_build/html/source/generated/pysegcnn.core.dataset.Cloud95Dataset.html
@@ -87,8 +87,8 @@
<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
<li class="toctree-l1 current"><a class="reference internal" href="../api.html">API Reference</a><ul class="current">
<li class="toctree-l2 current"><a class="reference internal" href="../api.html#dataset">Dataset</a><ul class="current">
-<li class="toctree-l3"><a class="reference internal" href="pysegcnn.core.dataset.ImageDataset.html">pysegcnn.core.dataset.ImageDataset</a></li>
-<li class="toctree-l3 current"><a class="reference internal" href="../api.html#supported-datasets">Supported datasets</a><ul class="current">
+<li class="toctree-l3"><a class="reference internal" href="../api.html#generic-classes">Generic classes</a></li>
+<li class="toctree-l3 current"><a class="reference internal" href="../api.html#specific-classes">Specific classes</a><ul class="current">
<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.dataset.SparcsDataset.html">pysegcnn.core.dataset.SparcsDataset</a></li>
<li class="toctree-l4 current"><a class="current reference internal" href="#">pysegcnn.core.dataset.Cloud95Dataset</a></li>
</ul>
@@ -171,80 +171,48 @@
<dt id="pysegcnn.core.dataset.Cloud95Dataset">
<em class="property">class </em><code class="sig-prename descclassname">pysegcnn.core.dataset.</code><code class="sig-name descname">Cloud95Dataset</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">root_dir</span></em>, <em class="sig-param"><span class="n">use_bands</span><span class="o">=</span><span class="default_value">[]</span></em>, <em class="sig-param"><span class="n">tile_size</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="n">pad</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">gt_pattern</span><span class="o">=</span><span class="default_value">'(.*)gt\\.tif'</span></em>, <em class="sig-param"><span class="n">sort</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">seed</span><span class="o">=</span><span class="default_value">0</span></em>, <em class="sig-param"><span class="n">transforms</span><span class="o">=</span><span class="default_value">[]</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/dataset.html#Cloud95Dataset"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.dataset.Cloud95Dataset" title="Permalink to this definition">¶</a></dt>
<dd><p>Class for the <a class="reference external" href="https://github.com/SorourMo/95-Cloud-An-Extension-to-38-Cloud-Dataset">Cloud-95</a> dataset by <a class="reference external" href="https://arxiv.org/abs/2001.08768">Mohajerani & Saeedi (2020)</a>.</p>
-<dl class="field-list simple">
+<dl class="py method">
+<dt id="pysegcnn.core.dataset.Cloud95Dataset.__init__">
+<code class="sig-name descname">__init__</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">root_dir</span></em>, <em class="sig-param"><span class="n">use_bands</span><span class="o">=</span><span class="default_value">[]</span></em>, <em class="sig-param"><span class="n">tile_size</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="n">pad</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">gt_pattern</span><span class="o">=</span><span class="default_value">'(.*)gt\\.tif'</span></em>, <em class="sig-param"><span class="n">sort</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">seed</span><span class="o">=</span><span class="default_value">0</span></em>, <em class="sig-param"><span class="n">transforms</span><span class="o">=</span><span class="default_value">[]</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/dataset.html#Cloud95Dataset.__init__"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.dataset.Cloud95Dataset.__init__" title="Permalink to this definition">¶</a></dt>
+<dd><p>Initialize.</p>
+<dl class="field-list">
<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>root_dir</strong> (<cite>str</cite>) – The root directory, path to the dataset.</p></li>
-<li><p><strong>use_bands</strong> (<cite>list</cite> [<cite>str</cite>], optional) – A list of the spectral bands to use. The default is [].</p></li>
-<li><p><strong>tile_size</strong> (<cite>int</cite> or <cite>None</cite>, optional) – The size of the tiles. If not <cite>None</cite>, each scene is divided into square
-tiles of shape (tile_size, tile_size). The default is None.</p></li>
-<li><p><strong>pad</strong> (<cite>bool</cite>, optional) – Whether to center pad the input image. Set <code class="docutils literal notranslate"><span class="pre">pad</span></code> = True, if the
-images are not evenly divisible by the <code class="docutils literal notranslate"><span class="pre">tile_size</span></code>. The image data is
-padded with a constant padding value of zero. For each image, the
-corresponding ground truth image is padded with a “no data†label.
-The default is False.</p></li>
-<li><p><strong>gt_pattern</strong> (<cite>str</cite>, optional) – A regural expression to match the ground truth naming convention. All
-directories and subdirectories in <code class="docutils literal notranslate"><span class="pre">root_dir</span></code> are searched for files
-matching <code class="docutils literal notranslate"><span class="pre">gt_pattern</span></code>. The default is ‘(.*)gt\.tif’.</p></li>
-<li><p><strong>sort</strong> (<cite>bool</cite>, optional) – Whether to chronologically sort the samples. Useful for time series
-data. The default is False.</p></li>
-<li><p><strong>seed</strong> (<cite>int</cite>, optional) – The random seed. Used to split the dataset into training, validation
-and test set. Useful for reproducibility. The default is 0.</p></li>
-<li><p><strong>transforms</strong> (<cite>list</cite> [<cite>pysegcnn.core.split.Augment</cite>], optional) – List of <cite>pysegcnn.core.split.Augment</cite> instances. Each item in
-<code class="docutils literal notranslate"><span class="pre">transforms</span></code> generates a distinct transformed version of the dataset.
-The total dataset is composed of the original untransformed dataset
-together with each transformed version of it.
-If <code class="docutils literal notranslate"><span class="pre">transforms</span></code> = [], only the original dataset is used.
-The default is [].</p></li>
-</ul>
+<dd class="field-odd"><dl>
+<dt><strong>root_dir</strong><span class="classifier"><cite>str</cite></span></dt><dd><p>The root directory, path to the dataset.</p>
+</dd>
+<dt><strong>use_bands</strong><span class="classifier"><cite>list</cite> [<cite>str</cite>], optional</span></dt><dd><p>A list of the spectral bands to use. The default is <cite>[]</cite>.</p>
</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
+<dt><strong>tile_size</strong><span class="classifier"><cite>int</cite> or <cite>None</cite>, optional</span></dt><dd><p>The size of the tiles. If not <cite>None</cite>, each scene is divided into
+square tiles of shape <code class="docutils literal notranslate"><span class="pre">(tile_size,</span> <span class="pre">tile_size)</span></code>. The default is
+<cite>None</cite>.</p>
</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
+<dt><strong>pad</strong><span class="classifier"><cite>bool</cite>, optional</span></dt><dd><p>Whether to center pad the input image. Set <code class="docutils literal notranslate"><span class="pre">pad=True</span></code>, if the
+images are not evenly divisible by the <code class="docutils literal notranslate"><span class="pre">tile_size</span></code>. The image
+data is padded with a constant padding value of zero. For each
+image, the corresponding ground truth image is padded with a
+“no data†label. The default is <cite>False</cite>.</p>
+</dd>
+<dt><strong>gt_pattern</strong><span class="classifier"><cite>str</cite>, optional</span></dt><dd><p>A regural expression to match the ground truth naming convention.
+All directories and subdirectories in <code class="docutils literal notranslate"><span class="pre">root_dir</span></code> are searched for
+files matching <code class="docutils literal notranslate"><span class="pre">gt_pattern</span></code>. The default is <cite>(.*)gt\.tif</cite>.</p>
+</dd>
+<dt><strong>sort</strong><span class="classifier"><cite>bool</cite>, optional</span></dt><dd><p>Whether to chronologically sort the samples. Useful for time series
+data. The default is <cite>False</cite>.</p>
+</dd>
+<dt><strong>seed</strong><span class="classifier"><cite>int</cite>, optional</span></dt><dd><p>The random seed. Used to split the dataset into training,
+validation and test set. Useful for reproducibility. The default is
+<cite>0</cite>.</p>
+</dd>
+<dt><strong>transforms</strong><span class="classifier"><cite>list</cite>, optional</span></dt><dd><p>List of <a class="reference internal" href="pysegcnn.core.transforms.Augment.html#pysegcnn.core.transforms.Augment" title="pysegcnn.core.transforms.Augment"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.transforms.Augment</span></code></a> instances.
+Each item in <code class="docutils literal notranslate"><span class="pre">transforms</span></code> generates a distinct transformed
+version of the dataset. The total dataset is composed of the
+original untransformed dataset together with each transformed
+version of it. If <code class="docutils literal notranslate"><span class="pre">transforms=[]</span></code>, only the original dataset is
+used. The default is <cite>[]</cite>.</p>
+</dd>
+</dl>
</dd>
</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">build_samples</span></code>(scene)</p></td>
-<td><p>Stack the bands of a sample in a single array.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">compose_scenes</span></code>()</p></td>
-<td><p>Build the list of samples of the dataset.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_labels</span></code>()</p></td>
-<td><p>Class labels of the Cloud-95 dataset.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_sensor</span></code>()</p></td>
-<td><p>Landsat 8 bands of the Cloud-95 dataset.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_size</span></code>()</p></td>
-<td><p>Image size of the Cloud-95 dataset.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parse_scene_id</span></code>(scene_id)</p></td>
-<td><p>Parse Sparcs scene identifiers (Landsat 8).</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">preprocess</span></code>(data, gt)</p></td>
-<td><p>Preprocess Cloud-95 dataset images.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">read_scene</span></code>(idx)</p></td>
-<td><p>Read the data of the sample with index <code class="docutils literal notranslate"><span class="pre">idx</span></code>.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to_tensor</span></code>(x, dtype)</p></td>
-<td><p>Convert <code class="docutils literal notranslate"><span class="pre">x</span></code> to <cite>torch.Tensor</cite>.</p></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="pysegcnn.core.dataset.Cloud95Dataset.__init__">
-<code class="sig-name descname">__init__</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">root_dir</span></em>, <em class="sig-param"><span class="n">use_bands</span><span class="o">=</span><span class="default_value">[]</span></em>, <em class="sig-param"><span class="n">tile_size</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="n">pad</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">gt_pattern</span><span class="o">=</span><span class="default_value">'(.*)gt\\.tif'</span></em>, <em class="sig-param"><span class="n">sort</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">seed</span><span class="o">=</span><span class="default_value">0</span></em>, <em class="sig-param"><span class="n">transforms</span><span class="o">=</span><span class="default_value">[]</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/dataset.html#Cloud95Dataset.__init__"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.dataset.Cloud95Dataset.__init__" title="Permalink to this definition">¶</a></dt>
-<dd><p>Initialize self. See help(type(self)) for accurate signature.</p>
</dd></dl>
<p class="rubric">Methods</p>
@@ -255,7 +223,7 @@ The default is [].</p></li>
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#pysegcnn.core.dataset.Cloud95Dataset.__init__" title="pysegcnn.core.dataset.Cloud95Dataset.__init__"><code class="xref py py-obj docutils literal notranslate"><span class="pre">__init__</span></code></a>(root_dir[, use_bands, tile_size, …])</p></td>
-<td><p>Initialize self.</p></td>
+<td><p>Initialize.</p></td>
</tr>
<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">build_samples</span></code>(scene)</p></td>
<td><p>Stack the bands of a sample in a single array.</p></td>
@@ -282,7 +250,7 @@ The default is [].</p></li>
<td><p>Read the data of the sample with index <code class="docutils literal notranslate"><span class="pre">idx</span></code>.</p></td>
</tr>
<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to_tensor</span></code>(x, dtype)</p></td>
-<td><p>Convert <code class="docutils literal notranslate"><span class="pre">x</span></code> to <cite>torch.Tensor</cite>.</p></td>
+<td><p>Convert <code class="docutils literal notranslate"><span class="pre">x</span></code> to <code class="xref py py-class docutils literal notranslate"><span class="pre">torch.Tensor</span></code>.</p></td>
</tr>
</tbody>
</table>
diff --git a/docs/_build/html/source/generated/pysegcnn.core.dataset.ImageDataset.html b/docs/_build/html/source/generated/pysegcnn.core.dataset.ImageDataset.html
index 31e363c90e798c8ed4afc43b8acb2964faf25e18..6ee1bfb83bd7c91b22ac749927f6d6ab8c58c8b4 100644
--- a/docs/_build/html/source/generated/pysegcnn.core.dataset.ImageDataset.html
+++ b/docs/_build/html/source/generated/pysegcnn.core.dataset.ImageDataset.html
@@ -36,7 +36,7 @@
<link rel="index" title="Index" href="../../genindex.html" />
<link rel="search" title="Search" href="../../search.html" />
- <link rel="next" title="pysegcnn.core.dataset.SparcsDataset" href="pysegcnn.core.dataset.SparcsDataset.html" />
+ <link rel="next" title="pysegcnn.core.dataset.StandardEoDataset" href="pysegcnn.core.dataset.StandardEoDataset.html" />
<link rel="prev" title="API Reference" href="../api.html" />
</head>
@@ -87,8 +87,12 @@
<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
<li class="toctree-l1 current"><a class="reference internal" href="../api.html">API Reference</a><ul class="current">
<li class="toctree-l2 current"><a class="reference internal" href="../api.html#dataset">Dataset</a><ul class="current">
-<li class="toctree-l3 current"><a class="current reference internal" href="#">pysegcnn.core.dataset.ImageDataset</a></li>
-<li class="toctree-l3"><a class="reference internal" href="../api.html#supported-datasets">Supported datasets</a></li>
+<li class="toctree-l3 current"><a class="reference internal" href="../api.html#generic-classes">Generic classes</a><ul class="current">
+<li class="toctree-l4 current"><a class="current reference internal" href="#">pysegcnn.core.dataset.ImageDataset</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.dataset.StandardEoDataset.html">pysegcnn.core.dataset.StandardEoDataset</a></li>
+</ul>
+</li>
+<li class="toctree-l3"><a class="reference internal" href="../api.html#specific-classes">Specific classes</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="../api.html#models">Models</a></li>
@@ -167,84 +171,51 @@
<dt id="pysegcnn.core.dataset.ImageDataset">
<em class="property">class </em><code class="sig-prename descclassname">pysegcnn.core.dataset.</code><code class="sig-name descname">ImageDataset</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">root_dir</span></em>, <em class="sig-param"><span class="n">use_bands</span><span class="o">=</span><span class="default_value">[]</span></em>, <em class="sig-param"><span class="n">tile_size</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="n">pad</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">gt_pattern</span><span class="o">=</span><span class="default_value">'(.*)gt.tif'</span></em>, <em class="sig-param"><span class="n">sort</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">seed</span><span class="o">=</span><span class="default_value">0</span></em>, <em class="sig-param"><span class="n">transforms</span><span class="o">=</span><span class="default_value">[]</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/dataset.html#ImageDataset"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.dataset.ImageDataset" title="Permalink to this definition">¶</a></dt>
<dd><p>Base class for multispectral image data.</p>
-<p>Inheriting from <cite>torch.utils.data.Dataset</cite> to be compliant to the PyTorch
-standard. Furthermore, using instances of <cite>torch.utils.data.Dataset</cite>
-enables the use of the handy <cite>torch.utils.data.DataLoader</cite> class during
-model training.</p>
-<dl class="field-list simple">
+<p>Inheriting from <code class="xref py py-class docutils literal notranslate"><span class="pre">torch.utils.data.Dataset</span></code> to be compliant to the
+PyTorch standard. This enables the use of the handy
+<code class="xref py py-class docutils literal notranslate"><span class="pre">torch.utils.data.DataLoader</span></code> class during model training.</p>
+<dl class="py method">
+<dt id="pysegcnn.core.dataset.ImageDataset.__init__">
+<code class="sig-name descname">__init__</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">root_dir</span></em>, <em class="sig-param"><span class="n">use_bands</span><span class="o">=</span><span class="default_value">[]</span></em>, <em class="sig-param"><span class="n">tile_size</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="n">pad</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">gt_pattern</span><span class="o">=</span><span class="default_value">'(.*)gt.tif'</span></em>, <em class="sig-param"><span class="n">sort</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">seed</span><span class="o">=</span><span class="default_value">0</span></em>, <em class="sig-param"><span class="n">transforms</span><span class="o">=</span><span class="default_value">[]</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/dataset.html#ImageDataset.__init__"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.dataset.ImageDataset.__init__" title="Permalink to this definition">¶</a></dt>
+<dd><p>Initialize.</p>
+<dl class="field-list">
<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>root_dir</strong> (<cite>str</cite>) – The root directory, path to the dataset.</p></li>
-<li><p><strong>use_bands</strong> (<cite>list</cite> [<cite>str</cite>], optional) – A list of the spectral bands to use. The default is [].</p></li>
-<li><p><strong>tile_size</strong> (<cite>int</cite> or <cite>None</cite>, optional) – The size of the tiles. If not <cite>None</cite>, each scene is divided into square
-tiles of shape (tile_size, tile_size). The default is None.</p></li>
-<li><p><strong>pad</strong> (<cite>bool</cite>, optional) – Whether to center pad the input image. Set <code class="docutils literal notranslate"><span class="pre">pad</span></code> = True, if the
-images are not evenly divisible by the <code class="docutils literal notranslate"><span class="pre">tile_size</span></code>. The image data is
-padded with a constant padding value of zero. For each image, the
-corresponding ground truth image is padded with a “no data†label.
-The default is False.</p></li>
-<li><p><strong>gt_pattern</strong> (<cite>str</cite>, optional) – A regural expression to match the ground truth naming convention.
+<dd class="field-odd"><dl>
+<dt><strong>root_dir</strong><span class="classifier"><cite>str</cite></span></dt><dd><p>The root directory, path to the dataset.</p>
+</dd>
+<dt><strong>use_bands</strong><span class="classifier"><cite>list</cite> [<cite>str</cite>], optional</span></dt><dd><p>A list of the spectral bands to use. The default is <cite>[]</cite>.</p>
+</dd>
+<dt><strong>tile_size</strong><span class="classifier"><cite>int</cite> or <cite>None</cite>, optional</span></dt><dd><p>The size of the tiles. If not <cite>None</cite>, each scene is divided into
+square tiles of shape <code class="docutils literal notranslate"><span class="pre">(tile_size,</span> <span class="pre">tile_size)</span></code>. The default is
+<cite>None</cite>.</p>
+</dd>
+<dt><strong>pad</strong><span class="classifier"><cite>bool</cite>, optional</span></dt><dd><p>Whether to center pad the input image. Set <code class="docutils literal notranslate"><span class="pre">pad=True</span></code>, if the
+images are not evenly divisible by the <code class="docutils literal notranslate"><span class="pre">tile_size</span></code>. The image
+data is padded with a constant padding value of zero. For each
+image, the corresponding ground truth image is padded with a
+“no data†label. The default is <cite>False</cite>.</p>
+</dd>
+<dt><strong>gt_pattern</strong><span class="classifier"><cite>str</cite>, optional</span></dt><dd><p>A regural expression to match the ground truth naming convention.
All directories and subdirectories in <code class="docutils literal notranslate"><span class="pre">root_dir</span></code> are searched for
-files matching <code class="docutils literal notranslate"><span class="pre">gt_pattern</span></code>. The default is ‘(.*)gt\.tif’.</p></li>
-<li><p><strong>sort</strong> (<cite>bool</cite>, optional) – Whether to chronologically sort the samples. Useful for time series
-data. The default is False.</p></li>
-<li><p><strong>seed</strong> (<cite>int</cite>, optional) – The random seed. Used to split the dataset into training, validation
-and test set. Useful for reproducibility. The default is 0.</p></li>
-<li><p><strong>transforms</strong> (<cite>list</cite> [<cite>pysegcnn.core.split.Augment</cite>], optional) – List of <cite>pysegcnn.core.split.Augment</cite> instances. Each item in
-<code class="docutils literal notranslate"><span class="pre">transforms</span></code> generates a distinct transformed version of the dataset.
-The total dataset is composed of the original untransformed dataset
-together with each transformed version of it.
-If <code class="docutils literal notranslate"><span class="pre">transforms</span></code> = [], only the original dataset is used.
-The default is [].</p></li>
-</ul>
+files matching <code class="docutils literal notranslate"><span class="pre">gt_pattern</span></code>. The default is <cite>(.*)gt\.tif</cite>.</p>
+</dd>
+<dt><strong>sort</strong><span class="classifier"><cite>bool</cite>, optional</span></dt><dd><p>Whether to chronologically sort the samples. Useful for time series
+data. The default is <cite>False</cite>.</p>
</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
+<dt><strong>seed</strong><span class="classifier"><cite>int</cite>, optional</span></dt><dd><p>The random seed. Used to split the dataset into training,
+validation and test set. Useful for reproducibility. The default is
+<cite>0</cite>.</p>
</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
+<dt><strong>transforms</strong><span class="classifier"><cite>list</cite>, optional</span></dt><dd><p>List of <a class="reference internal" href="pysegcnn.core.transforms.Augment.html#pysegcnn.core.transforms.Augment" title="pysegcnn.core.transforms.Augment"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.transforms.Augment</span></code></a> instances.
+Each item in <code class="docutils literal notranslate"><span class="pre">transforms</span></code> generates a distinct transformed
+version of the dataset. The total dataset is composed of the
+original untransformed dataset together with each transformed
+version of it. If <code class="docutils literal notranslate"><span class="pre">transforms=[]</span></code>, only the original dataset is
+used. The default is <cite>[]</cite>.</p>
+</dd>
+</dl>
</dd>
</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">build_samples</span></code>(scene)</p></td>
-<td><p>Stack the bands of a sample in a single array.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">compose_scenes</span></code>()</p></td>
-<td><p>Build the list of samples of the dataset.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_labels</span></code>()</p></td>
-<td><p>Return an enumeration of the class labels of the dataset.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_sensor</span></code>()</p></td>
-<td><p>Return an enumeration of the bands of the sensor of the dataset.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_size</span></code>()</p></td>
-<td><p>Return the size of the images in the dataset.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parse_scene_id</span></code>(scene_id)</p></td>
-<td><p>Parse the scene identifier.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">preprocess</span></code>(data, gt)</p></td>
-<td><p>Preprocess a sample before feeding it to a model.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">read_scene</span></code>(idx)</p></td>
-<td><p>Read the data of the sample with index <code class="docutils literal notranslate"><span class="pre">idx</span></code>.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to_tensor</span></code>(x, dtype)</p></td>
-<td><p>Convert <code class="docutils literal notranslate"><span class="pre">x</span></code> to <cite>torch.Tensor</cite>.</p></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="pysegcnn.core.dataset.ImageDataset.__init__">
-<code class="sig-name descname">__init__</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">root_dir</span></em>, <em class="sig-param"><span class="n">use_bands</span><span class="o">=</span><span class="default_value">[]</span></em>, <em class="sig-param"><span class="n">tile_size</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="n">pad</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">gt_pattern</span><span class="o">=</span><span class="default_value">'(.*)gt.tif'</span></em>, <em class="sig-param"><span class="n">sort</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">seed</span><span class="o">=</span><span class="default_value">0</span></em>, <em class="sig-param"><span class="n">transforms</span><span class="o">=</span><span class="default_value">[]</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/dataset.html#ImageDataset.__init__"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.dataset.ImageDataset.__init__" title="Permalink to this definition">¶</a></dt>
-<dd><p>Initialize self. See help(type(self)) for accurate signature.</p>
</dd></dl>
<p class="rubric">Methods</p>
@@ -255,7 +226,7 @@ The default is [].</p></li>
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#pysegcnn.core.dataset.ImageDataset.__init__" title="pysegcnn.core.dataset.ImageDataset.__init__"><code class="xref py py-obj docutils literal notranslate"><span class="pre">__init__</span></code></a>(root_dir[, use_bands, tile_size, …])</p></td>
-<td><p>Initialize self.</p></td>
+<td><p>Initialize.</p></td>
</tr>
<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">build_samples</span></code>(scene)</p></td>
<td><p>Stack the bands of a sample in a single array.</p></td>
@@ -282,7 +253,7 @@ The default is [].</p></li>
<td><p>Read the data of the sample with index <code class="docutils literal notranslate"><span class="pre">idx</span></code>.</p></td>
</tr>
<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to_tensor</span></code>(x, dtype)</p></td>
-<td><p>Convert <code class="docutils literal notranslate"><span class="pre">x</span></code> to <cite>torch.Tensor</cite>.</p></td>
+<td><p>Convert <code class="docutils literal notranslate"><span class="pre">x</span></code> to <code class="xref py py-class docutils literal notranslate"><span class="pre">torch.Tensor</span></code>.</p></td>
</tr>
</tbody>
</table>
@@ -298,7 +269,7 @@ The default is [].</p></li>
<div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
- <a href="pysegcnn.core.dataset.SparcsDataset.html" class="btn btn-neutral float-right" title="pysegcnn.core.dataset.SparcsDataset" accesskey="n" rel="next">Next <span class="fa fa-arrow-circle-right"></span></a>
+ <a href="pysegcnn.core.dataset.StandardEoDataset.html" class="btn btn-neutral float-right" title="pysegcnn.core.dataset.StandardEoDataset" accesskey="n" rel="next">Next <span class="fa fa-arrow-circle-right"></span></a>
<a href="../api.html" class="btn btn-neutral float-left" title="API Reference" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left"></span> Previous</a>
diff --git a/docs/_build/html/source/generated/pysegcnn.core.dataset.SparcsDataset.html b/docs/_build/html/source/generated/pysegcnn.core.dataset.SparcsDataset.html
index 9aa65a8dd2f76b40cc0f1f8117cd184ec1677b6f..37b1630b6319393e6c0e06e9c93ee29ba373a3af 100644
--- a/docs/_build/html/source/generated/pysegcnn.core.dataset.SparcsDataset.html
+++ b/docs/_build/html/source/generated/pysegcnn.core.dataset.SparcsDataset.html
@@ -37,7 +37,7 @@
<link rel="index" title="Index" href="../../genindex.html" />
<link rel="search" title="Search" href="../../search.html" />
<link rel="next" title="pysegcnn.core.dataset.Cloud95Dataset" href="pysegcnn.core.dataset.Cloud95Dataset.html" />
- <link rel="prev" title="pysegcnn.core.dataset.ImageDataset" href="pysegcnn.core.dataset.ImageDataset.html" />
+ <link rel="prev" title="pysegcnn.core.dataset.StandardEoDataset" href="pysegcnn.core.dataset.StandardEoDataset.html" />
</head>
<body class="wy-body-for-nav">
@@ -87,8 +87,8 @@
<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
<li class="toctree-l1 current"><a class="reference internal" href="../api.html">API Reference</a><ul class="current">
<li class="toctree-l2 current"><a class="reference internal" href="../api.html#dataset">Dataset</a><ul class="current">
-<li class="toctree-l3"><a class="reference internal" href="pysegcnn.core.dataset.ImageDataset.html">pysegcnn.core.dataset.ImageDataset</a></li>
-<li class="toctree-l3 current"><a class="reference internal" href="../api.html#supported-datasets">Supported datasets</a><ul class="current">
+<li class="toctree-l3"><a class="reference internal" href="../api.html#generic-classes">Generic classes</a></li>
+<li class="toctree-l3 current"><a class="reference internal" href="../api.html#specific-classes">Specific classes</a><ul class="current">
<li class="toctree-l4 current"><a class="current reference internal" href="#">pysegcnn.core.dataset.SparcsDataset</a></li>
<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.dataset.Cloud95Dataset.html">pysegcnn.core.dataset.Cloud95Dataset</a></li>
</ul>
@@ -170,81 +170,49 @@
<dl class="py class">
<dt id="pysegcnn.core.dataset.SparcsDataset">
<em class="property">class </em><code class="sig-prename descclassname">pysegcnn.core.dataset.</code><code class="sig-name descname">SparcsDataset</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">root_dir</span></em>, <em class="sig-param"><span class="n">use_bands</span><span class="o">=</span><span class="default_value">[]</span></em>, <em class="sig-param"><span class="n">tile_size</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="n">pad</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">gt_pattern</span><span class="o">=</span><span class="default_value">'(.*)gt\\.tif'</span></em>, <em class="sig-param"><span class="n">sort</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">seed</span><span class="o">=</span><span class="default_value">0</span></em>, <em class="sig-param"><span class="n">transforms</span><span class="o">=</span><span class="default_value">[]</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/dataset.html#SparcsDataset"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.dataset.SparcsDataset" title="Permalink to this definition">¶</a></dt>
-<dd><p>Class for the <a class="reference external" href="https://www.usgs.gov/land-resources/nli/landsat/spatial-procedures-automated-removal-cloud-and-shadow-sparcs-validation">Sparcs</a> dataset.</p>
-<dl class="field-list simple">
+<dd><p>Class for the <a class="reference external" href="https://www.usgs.gov/land-resources/nli/landsat/spatial-procedures-automated-removal-cloud-and-shadow-sparcs-validation">Sparcs</a> dataset by <a class="reference external" href="https://www.mdpi.com/2072-4292/6/6/4907">Hughes & Hayes (2014)</a>.</p>
+<dl class="py method">
+<dt id="pysegcnn.core.dataset.SparcsDataset.__init__">
+<code class="sig-name descname">__init__</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">root_dir</span></em>, <em class="sig-param"><span class="n">use_bands</span><span class="o">=</span><span class="default_value">[]</span></em>, <em class="sig-param"><span class="n">tile_size</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="n">pad</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">gt_pattern</span><span class="o">=</span><span class="default_value">'(.*)gt\\.tif'</span></em>, <em class="sig-param"><span class="n">sort</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">seed</span><span class="o">=</span><span class="default_value">0</span></em>, <em class="sig-param"><span class="n">transforms</span><span class="o">=</span><span class="default_value">[]</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/dataset.html#SparcsDataset.__init__"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.dataset.SparcsDataset.__init__" title="Permalink to this definition">¶</a></dt>
+<dd><p>Initialize.</p>
+<dl class="field-list">
<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>root_dir</strong> (<cite>str</cite>) – The root directory, path to the dataset.</p></li>
-<li><p><strong>use_bands</strong> (<cite>list</cite> [<cite>str</cite>], optional) – A list of the spectral bands to use. The default is [].</p></li>
-<li><p><strong>tile_size</strong> (<cite>int</cite> or <cite>None</cite>, optional) – The size of the tiles. If not <cite>None</cite>, each scene is divided into square
-tiles of shape (tile_size, tile_size). The default is None.</p></li>
-<li><p><strong>pad</strong> (<cite>bool</cite>, optional) – Whether to center pad the input image. Set <code class="docutils literal notranslate"><span class="pre">pad</span></code> = True, if the
-images are not evenly divisible by the <code class="docutils literal notranslate"><span class="pre">tile_size</span></code>. The image data is
-padded with a constant padding value of zero. For each image, the
-corresponding ground truth image is padded with a “no data†label.
-The default is False.</p></li>
-<li><p><strong>gt_pattern</strong> (<cite>str</cite>, optional) – A regural expression to match the ground truth naming convention. All
-directories and subdirectories in <code class="docutils literal notranslate"><span class="pre">root_dir</span></code> are searched for files
-matching <code class="docutils literal notranslate"><span class="pre">gt_pattern</span></code>. The default is ‘(.*)gt\.tif’.</p></li>
-<li><p><strong>sort</strong> (<cite>bool</cite>, optional) – Whether to chronologically sort the samples. Useful for time series
-data. The default is False.</p></li>
-<li><p><strong>seed</strong> (<cite>int</cite>, optional) – The random seed. Used to split the dataset into training, validation
-and test set. Useful for reproducibility. The default is 0.</p></li>
-<li><p><strong>transforms</strong> (<cite>list</cite> [<cite>pysegcnn.core.split.Augment</cite>], optional) – List of <cite>pysegcnn.core.split.Augment</cite> instances. Each item in
-<code class="docutils literal notranslate"><span class="pre">transforms</span></code> generates a distinct transformed version of the dataset.
-The total dataset is composed of the original untransformed dataset
-together with each transformed version of it.
-If <code class="docutils literal notranslate"><span class="pre">transforms</span></code> = [], only the original dataset is used.
-The default is [].</p></li>
-</ul>
+<dd class="field-odd"><dl>
+<dt><strong>root_dir</strong><span class="classifier"><cite>str</cite></span></dt><dd><p>The root directory, path to the dataset.</p>
+</dd>
+<dt><strong>use_bands</strong><span class="classifier"><cite>list</cite> [<cite>str</cite>], optional</span></dt><dd><p>A list of the spectral bands to use. The default is <cite>[]</cite>.</p>
</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
+<dt><strong>tile_size</strong><span class="classifier"><cite>int</cite> or <cite>None</cite>, optional</span></dt><dd><p>The size of the tiles. If not <cite>None</cite>, each scene is divided into
+square tiles of shape <code class="docutils literal notranslate"><span class="pre">(tile_size,</span> <span class="pre">tile_size)</span></code>. The default is
+<cite>None</cite>.</p>
</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
+<dt><strong>pad</strong><span class="classifier"><cite>bool</cite>, optional</span></dt><dd><p>Whether to center pad the input image. Set <code class="docutils literal notranslate"><span class="pre">pad=True</span></code>, if the
+images are not evenly divisible by the <code class="docutils literal notranslate"><span class="pre">tile_size</span></code>. The image
+data is padded with a constant padding value of zero. For each
+image, the corresponding ground truth image is padded with a
+“no data†label. The default is <cite>False</cite>.</p>
+</dd>
+<dt><strong>gt_pattern</strong><span class="classifier"><cite>str</cite>, optional</span></dt><dd><p>A regural expression to match the ground truth naming convention.
+All directories and subdirectories in <code class="docutils literal notranslate"><span class="pre">root_dir</span></code> are searched for
+files matching <code class="docutils literal notranslate"><span class="pre">gt_pattern</span></code>. The default is <cite>(.*)gt\.tif</cite>.</p>
+</dd>
+<dt><strong>sort</strong><span class="classifier"><cite>bool</cite>, optional</span></dt><dd><p>Whether to chronologically sort the samples. Useful for time series
+data. The default is <cite>False</cite>.</p>
+</dd>
+<dt><strong>seed</strong><span class="classifier"><cite>int</cite>, optional</span></dt><dd><p>The random seed. Used to split the dataset into training,
+validation and test set. Useful for reproducibility. The default is
+<cite>0</cite>.</p>
+</dd>
+<dt><strong>transforms</strong><span class="classifier"><cite>list</cite>, optional</span></dt><dd><p>List of <a class="reference internal" href="pysegcnn.core.transforms.Augment.html#pysegcnn.core.transforms.Augment" title="pysegcnn.core.transforms.Augment"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.transforms.Augment</span></code></a> instances.
+Each item in <code class="docutils literal notranslate"><span class="pre">transforms</span></code> generates a distinct transformed
+version of the dataset. The total dataset is composed of the
+original untransformed dataset together with each transformed
+version of it. If <code class="docutils literal notranslate"><span class="pre">transforms=[]</span></code>, only the original dataset is
+used. The default is <cite>[]</cite>.</p>
+</dd>
+</dl>
</dd>
</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">build_samples</span></code>(scene)</p></td>
-<td><p>Stack the bands of a sample in a single array.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">compose_scenes</span></code>()</p></td>
-<td><p>Build the list of samples of the dataset.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_labels</span></code>()</p></td>
-<td><p>Class labels of the Sparcs dataset.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_sensor</span></code>()</p></td>
-<td><p>Landsat 8 bands of the Sparcs dataset.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_size</span></code>()</p></td>
-<td><p>Image size of the Sparcs dataset.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parse_scene_id</span></code>(scene_id)</p></td>
-<td><p>Parse Sparcs scene identifiers (Landsat 8).</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">preprocess</span></code>(data, gt)</p></td>
-<td><p>Preprocess Sparcs dataset images.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">read_scene</span></code>(idx)</p></td>
-<td><p>Read the data of the sample with index <code class="docutils literal notranslate"><span class="pre">idx</span></code>.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to_tensor</span></code>(x, dtype)</p></td>
-<td><p>Convert <code class="docutils literal notranslate"><span class="pre">x</span></code> to <cite>torch.Tensor</cite>.</p></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="pysegcnn.core.dataset.SparcsDataset.__init__">
-<code class="sig-name descname">__init__</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">root_dir</span></em>, <em class="sig-param"><span class="n">use_bands</span><span class="o">=</span><span class="default_value">[]</span></em>, <em class="sig-param"><span class="n">tile_size</span><span class="o">=</span><span class="default_value">None</span></em>, <em class="sig-param"><span class="n">pad</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">gt_pattern</span><span class="o">=</span><span class="default_value">'(.*)gt\\.tif'</span></em>, <em class="sig-param"><span class="n">sort</span><span class="o">=</span><span class="default_value">False</span></em>, <em class="sig-param"><span class="n">seed</span><span class="o">=</span><span class="default_value">0</span></em>, <em class="sig-param"><span class="n">transforms</span><span class="o">=</span><span class="default_value">[]</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/dataset.html#SparcsDataset.__init__"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.dataset.SparcsDataset.__init__" title="Permalink to this definition">¶</a></dt>
-<dd><p>Initialize self. See help(type(self)) for accurate signature.</p>
</dd></dl>
<p class="rubric">Methods</p>
@@ -255,7 +223,7 @@ The default is [].</p></li>
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#pysegcnn.core.dataset.SparcsDataset.__init__" title="pysegcnn.core.dataset.SparcsDataset.__init__"><code class="xref py py-obj docutils literal notranslate"><span class="pre">__init__</span></code></a>(root_dir[, use_bands, tile_size, …])</p></td>
-<td><p>Initialize self.</p></td>
+<td><p>Initialize.</p></td>
</tr>
<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">build_samples</span></code>(scene)</p></td>
<td><p>Stack the bands of a sample in a single array.</p></td>
@@ -282,7 +250,7 @@ The default is [].</p></li>
<td><p>Read the data of the sample with index <code class="docutils literal notranslate"><span class="pre">idx</span></code>.</p></td>
</tr>
<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to_tensor</span></code>(x, dtype)</p></td>
-<td><p>Convert <code class="docutils literal notranslate"><span class="pre">x</span></code> to <cite>torch.Tensor</cite>.</p></td>
+<td><p>Convert <code class="docutils literal notranslate"><span class="pre">x</span></code> to <code class="xref py py-class docutils literal notranslate"><span class="pre">torch.Tensor</span></code>.</p></td>
</tr>
</tbody>
</table>
@@ -301,7 +269,7 @@ The default is [].</p></li>
<a href="pysegcnn.core.dataset.Cloud95Dataset.html" class="btn btn-neutral float-right" title="pysegcnn.core.dataset.Cloud95Dataset" accesskey="n" rel="next">Next <span class="fa fa-arrow-circle-right"></span></a>
- <a href="pysegcnn.core.dataset.ImageDataset.html" class="btn btn-neutral float-left" title="pysegcnn.core.dataset.ImageDataset" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left"></span> Previous</a>
+ <a href="pysegcnn.core.dataset.StandardEoDataset.html" class="btn btn-neutral float-left" title="pysegcnn.core.dataset.StandardEoDataset" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left"></span> Previous</a>
</div>
diff --git a/docs/_build/html/source/generated/pysegcnn.core.layers.Block.html b/docs/_build/html/source/generated/pysegcnn.core.layers.Block.html
index aca8887c887829393ef959bdc9a865f0fae8f672..a0aa79226b3ef4f428d355717407f42f2a3fa9e6 100644
--- a/docs/_build/html/source/generated/pysegcnn.core.layers.Block.html
+++ b/docs/_build/html/source/generated/pysegcnn.core.layers.Block.html
@@ -35,7 +35,9 @@
<link rel="index" title="Index" href="../../genindex.html" />
- <link rel="search" title="Search" href="../../search.html" />
+ <link rel="search" title="Search" href="../../search.html" />
+ <link rel="next" title="pysegcnn.core.layers.Conv2dSame" href="pysegcnn.core.layers.Conv2dSame.html" />
+ <link rel="prev" title="pysegcnn.core.dataset.Cloud95Dataset" href="pysegcnn.core.dataset.Cloud95Dataset.html" />
</head>
<body class="wy-body-for-nav">
@@ -80,10 +82,26 @@
- <ul>
+ <ul class="current">
<li class="toctree-l1"><a class="reference internal" href="../requirements.html">Requirements</a></li>
<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../api.html">API Reference</a></li>
+<li class="toctree-l1 current"><a class="reference internal" href="../api.html">API Reference</a><ul class="current">
+<li class="toctree-l2"><a class="reference internal" href="../api.html#dataset">Dataset</a></li>
+<li class="toctree-l2 current"><a class="reference internal" href="../api.html#models">Models</a><ul class="current">
+<li class="toctree-l3 current"><a class="reference internal" href="../api.html#layers">Layers</a><ul class="current">
+<li class="toctree-l4 current"><a class="current reference internal" href="#">pysegcnn.core.layers.Block</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.Conv2dSame.html">pysegcnn.core.layers.Conv2dSame</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.ConvBnReluMaxPool.html">pysegcnn.core.layers.ConvBnReluMaxPool</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.ConvBnReluMaxUnpool.html">pysegcnn.core.layers.ConvBnReluMaxUnpool</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.ConvBnReluUpsample.html">pysegcnn.core.layers.ConvBnReluUpsample</a></li>
+</ul>
+</li>
+<li class="toctree-l3"><a class="reference internal" href="../api.html#encoder-decoder-architechture">Encoder-Decoder architechture</a></li>
+<li class="toctree-l3"><a class="reference internal" href="../api.html#neural-networks">Neural Networks</a></li>
+</ul>
+</li>
+</ul>
+</li>
</ul>
@@ -130,6 +148,8 @@
<li><a href="../../index.html" class="icon icon-home"></a> »</li>
+ <li><a href="../api.html">API Reference</a> »</li>
+
<li>pysegcnn.core.layers.Block</li>
@@ -155,148 +175,45 @@
<dt id="pysegcnn.core.layers.Block">
<em class="property">class </em><code class="sig-prename descclassname">pysegcnn.core.layers.</code><code class="sig-name descname">Block</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">out_channels</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/layers.html#Block"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.layers.Block" title="Permalink to this definition">¶</a></dt>
<dd><p>Basic convolutional block.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>in_channels</strong> (<cite>int</cite>) – Number of input channels.</p></li>
-<li><p><strong>out_channels</strong> (<cite>int</cite>) – Number of output channels.</p></li>
-<li><p><strong>**kwargs</strong> (‘dict’ [<cite>str</cite>]) – Additional arguments passed to <cite>pysegcnn.core.layers.Conv2dSame</cite>.</p></li>
-</ul>
+<dl class="field-list">
+<dt class="field-odd">Attributes</dt>
+<dd class="field-odd"><dl>
+<dt><strong>in_channels</strong><span class="classifier"><cite>int</cite></span></dt><dd><p>Number of input channels.</p>
</dd>
-<dt class="field-even">Raises</dt>
-<dd class="field-even"><p><strong>TypeError</strong> – Raised if <cite>~pysegcnn.core.layers.Block.layers</cite> method does not return
-an instance of <cite>torch.nn.Sequential</cite>.</p>
+<dt><strong>out_channels</strong><span class="classifier"><cite>int</cite></span></dt><dd><p>Number of output channels.</p>
</dd>
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p></p>
+<dt><strong>kwargs: `dict` [`str`]</strong></dt><dd><p>Additional arguments passed to
+<a class="reference internal" href="pysegcnn.core.layers.Conv2dSame.html#pysegcnn.core.layers.Conv2dSame" title="pysegcnn.core.layers.Conv2dSame"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.Conv2dSame</span></code></a>.</p>
</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p>None.</p>
+<dt><strong>conv</strong><span class="classifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">torch.nn.Sequential</span></code></span></dt><dd><p>The convolutional layers of the block.</p>
+</dd>
+</dl>
</dd>
</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
-<td><p>Adds a child module to the current module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">apply</span></code>(fn)</p></td>
-<td><p>Applies <code class="docutils literal notranslate"><span class="pre">fn</span></code> recursively to every submodule (as returned by <code class="docutils literal notranslate"><span class="pre">.children()</span></code>) as well as self.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">bfloat16</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">bfloat16</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">buffers</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cpu</span></code>()</p></td>
-<td><p>Moves all model parameters and buffers to the CPU.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cuda</span></code>([device])</p></td>
-<td><p>Moves all model parameters and buffers to the GPU.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">double</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">double</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">eval</span></code>()</p></td>
-<td><p>Sets the module in evaluation mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">extra_repr</span></code>()</p></td>
-<td><p>Set the extra representation of the module</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">float</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to float datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward</span></code>()</p></td>
-<td><p>Forward pass of the block.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">half</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">half</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">layers</span></code>()</p></td>
-<td><p>Define the layers of the block.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_state_dict</span></code>(state_dict[, strict])</p></td>
-<td><p>Copies parameters and buffers from <code class="xref py py-attr docutils literal notranslate"><span class="pre">state_dict</span></code> into this module and its descendants.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">modules</span></code>()</p></td>
-<td><p>Returns an iterator over all modules in the network.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_buffers</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_modules</span></code>([memo, prefix])</p></td>
-<td><p>Returns an iterator over all modules in the network, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_parameters</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parameters</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module parameters.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_backward_hook</span></code>(hook)</p></td>
-<td><p>Registers a backward hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_buffer</span></code>(name, tensor[, persistent])</p></td>
-<td><p>Adds a buffer to the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_pre_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward pre-hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_parameter</span></code>(name, param)</p></td>
-<td><p>Adds a parameter to the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">requires_grad_</span></code>([requires_grad])</p></td>
-<td><p>Change if autograd should record operations on parameters in this module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">state_dict</span></code>([destination, prefix, keep_vars])</p></td>
-<td><p>Returns a dictionary containing a whole state of the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to</span></code>(*args, **kwargs)</p></td>
-<td><p>Moves and/or casts the parameters and buffers.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">train</span></code>([mode])</p></td>
-<td><p>Sets the module in training mode.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">type</span></code>(dst_type)</p></td>
-<td><p>Casts all parameters and buffers to <code class="xref py py-attr docutils literal notranslate"><span class="pre">dst_type</span></code>.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">zero_grad</span></code>()</p></td>
-<td><p>Sets gradients of all model parameters to zero.</p></td>
-</tr>
-</tbody>
-</table>
-<table class="docutils align-default">
-<colgroup>
-<col style="width: 62%" />
-<col style="width: 38%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><strong>__call__</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-even"><td><p><strong>share_memory</strong></p></td>
-<td></td>
-</tr>
-</tbody>
-</table>
<dl class="py method">
<dt id="pysegcnn.core.layers.Block.__init__">
<code class="sig-name descname">__init__</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">out_channels</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/layers.html#Block.__init__"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.layers.Block.__init__" title="Permalink to this definition">¶</a></dt>
-<dd><p>Initializes internal Module state, shared by both nn.Module and ScriptModule.</p>
+<dd><p>Initialize.</p>
+<dl class="field-list">
+<dt class="field-odd">Parameters</dt>
+<dd class="field-odd"><dl>
+<dt><strong>in_channels</strong><span class="classifier"><cite>int</cite></span></dt><dd><p>Number of input channels.</p>
+</dd>
+<dt><strong>out_channels</strong><span class="classifier"><cite>int</cite></span></dt><dd><p>Number of output channels.</p>
+</dd>
+<dt><strong>**kwargs: `dict` [`str`]</strong></dt><dd><p>Additional arguments passed to
+<a class="reference internal" href="pysegcnn.core.layers.Conv2dSame.html#pysegcnn.core.layers.Conv2dSame" title="pysegcnn.core.layers.Conv2dSame"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.Conv2dSame</span></code></a>.</p>
+</dd>
+</dl>
+</dd>
+<dt class="field-even">Raises</dt>
+<dd class="field-even"><dl class="simple">
+<dt>TypeError</dt><dd><p>Raised if <code class="xref py py-meth docutils literal notranslate"><span class="pre">layers()</span></code> method does
+not return an instance of <code class="xref py py-class docutils literal notranslate"><span class="pre">torch.nn.Sequential</span></code>.</p>
+</dd>
+</dl>
+</dd>
+</dl>
</dd></dl>
<p class="rubric">Methods</p>
@@ -307,7 +224,7 @@ an instance of <cite>torch.nn.Sequential</cite>.</p>
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#pysegcnn.core.layers.Block.__init__" title="pysegcnn.core.layers.Block.__init__"><code class="xref py py-obj docutils literal notranslate"><span class="pre">__init__</span></code></a>(in_channels, out_channels, **kwargs)</p></td>
-<td><p>Initializes internal Module state, shared by both nn.Module and ScriptModule.</p></td>
+<td><p>Initialize.</p></td>
</tr>
<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
<td><p>Adds a child module to the current module.</p></td>
@@ -435,6 +352,15 @@ an instance of <cite>torch.nn.Sequential</cite>.</p>
</div>
<footer>
+ <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
+
+ <a href="pysegcnn.core.layers.Conv2dSame.html" class="btn btn-neutral float-right" title="pysegcnn.core.layers.Conv2dSame" accesskey="n" rel="next">Next <span class="fa fa-arrow-circle-right"></span></a>
+
+
+ <a href="pysegcnn.core.dataset.Cloud95Dataset.html" class="btn btn-neutral float-left" title="pysegcnn.core.dataset.Cloud95Dataset" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left"></span> Previous</a>
+
+ </div>
+
<hr/>
diff --git a/docs/_build/html/source/generated/pysegcnn.core.layers.Conv2dSame.html b/docs/_build/html/source/generated/pysegcnn.core.layers.Conv2dSame.html
index 48ec4f834446ecb9f1d19c8ddfb1941d4c13446c..52f526271b86e8c2c7c460b71b31d2d7a7e717d7 100644
--- a/docs/_build/html/source/generated/pysegcnn.core.layers.Conv2dSame.html
+++ b/docs/_build/html/source/generated/pysegcnn.core.layers.Conv2dSame.html
@@ -35,7 +35,9 @@
<link rel="index" title="Index" href="../../genindex.html" />
- <link rel="search" title="Search" href="../../search.html" />
+ <link rel="search" title="Search" href="../../search.html" />
+ <link rel="next" title="pysegcnn.core.layers.ConvBnReluMaxPool" href="pysegcnn.core.layers.ConvBnReluMaxPool.html" />
+ <link rel="prev" title="pysegcnn.core.layers.Block" href="pysegcnn.core.layers.Block.html" />
</head>
<body class="wy-body-for-nav">
@@ -80,10 +82,26 @@
- <ul>
+ <ul class="current">
<li class="toctree-l1"><a class="reference internal" href="../requirements.html">Requirements</a></li>
<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../api.html">API Reference</a></li>
+<li class="toctree-l1 current"><a class="reference internal" href="../api.html">API Reference</a><ul class="current">
+<li class="toctree-l2"><a class="reference internal" href="../api.html#dataset">Dataset</a></li>
+<li class="toctree-l2 current"><a class="reference internal" href="../api.html#models">Models</a><ul class="current">
+<li class="toctree-l3 current"><a class="reference internal" href="../api.html#layers">Layers</a><ul class="current">
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.Block.html">pysegcnn.core.layers.Block</a></li>
+<li class="toctree-l4 current"><a class="current reference internal" href="#">pysegcnn.core.layers.Conv2dSame</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.ConvBnReluMaxPool.html">pysegcnn.core.layers.ConvBnReluMaxPool</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.ConvBnReluMaxUnpool.html">pysegcnn.core.layers.ConvBnReluMaxUnpool</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.ConvBnReluUpsample.html">pysegcnn.core.layers.ConvBnReluUpsample</a></li>
+</ul>
+</li>
+<li class="toctree-l3"><a class="reference internal" href="../api.html#encoder-decoder-architechture">Encoder-Decoder architechture</a></li>
+<li class="toctree-l3"><a class="reference internal" href="../api.html#neural-networks">Neural Networks</a></li>
+</ul>
+</li>
+</ul>
+</li>
</ul>
@@ -130,6 +148,8 @@
<li><a href="../../index.html" class="icon icon-home"></a> »</li>
+ <li><a href="../api.html">API Reference</a> »</li>
+
<li>pysegcnn.core.layers.Conv2dSame</li>
@@ -158,11 +178,23 @@
<p>Given the kernel size, the dilation and a stride of 1, the padding is
calculated such that the output of the convolution has the same spatial
dimensions as the input.</p>
+<dl class="field-list">
+<dt class="field-odd">Attributes</dt>
+<dd class="field-odd"><dl>
+<dt><strong>padding</strong><span class="classifier"><cite>tuple</cite> [<cite>int</cite>]</span></dt><dd><p>The amount of padding, (pad_height, pad_width).</p>
+</dd>
+</dl>
+</dd>
+</dl>
+<dl class="py method">
+<dt id="pysegcnn.core.layers.Conv2dSame.__init__">
+<code class="sig-name descname">__init__</code><span class="sig-paren">(</span><em class="sig-param"><span class="o">*</span><span class="n">args</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/layers.html#Conv2dSame.__init__"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.layers.Conv2dSame.__init__" title="Permalink to this definition">¶</a></dt>
+<dd><p>Initialize.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>*args</strong> (<cite>list</cite> [<cite>str</cite>]) – <dl class="simple">
-<dt>positional arguments passed to <cite>torch.nn.Conv2d</cite>:</dt><dd><dl class="simple">
+<dd class="field-odd"><dl class="simple">
+<dt><strong>*args: `list` [`str`]</strong></dt><dd><dl class="simple">
+<dt>positional arguments passed to <code class="xref py py-class docutils literal notranslate"><span class="pre">torch.nn.Conv2d</span></code>:</dt><dd><dl class="simple">
<dt><code class="docutils literal notranslate"><span class="pre">'in_channels'</span></code>: <cite>int</cite></dt><dd><p>Number of input channels.</p>
</dd>
<dt><code class="docutils literal notranslate"><span class="pre">'out_channels'</span></code>: <cite>int</cite></dt><dd><p>Number of output channels.</p>
@@ -172,144 +204,12 @@ dimensions as the input.</p>
</dl>
</dd>
</dl>
-</p></li>
-<li><p><strong>**kwargs</strong> (‘dict’ [<cite>str</cite>]) – Additional keyword arguments passed to <a href="#id1"><span class="problematic" id="id2">`torch.nn.Conv2d`_</span></a>.</p></li>
-<li><p><strong>_torch.nn.Conv2d</strong> (<em>.</em>) – <a class="reference external" href="https://pytorch.org/docs/stable/generated/torch.nn.Conv2d.html#torch.nn.Conv2d">https://pytorch.org/docs/stable/generated/torch.nn.Conv2d.html#torch.nn.Conv2d</a></p></li>
-</ul>
</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
+<dt><strong>**kwargs: `dict` [`str`]</strong></dt><dd><p>Additional keyword arguments passed to <code class="xref py py-class docutils literal notranslate"><span class="pre">torch.nn.Conv2d</span></code>.</p>
</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
+</dl>
</dd>
</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
-<td><p>Adds a child module to the current module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">apply</span></code>(fn)</p></td>
-<td><p>Applies <code class="docutils literal notranslate"><span class="pre">fn</span></code> recursively to every submodule (as returned by <code class="docutils literal notranslate"><span class="pre">.children()</span></code>) as well as self.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">bfloat16</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">bfloat16</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">buffers</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cpu</span></code>()</p></td>
-<td><p>Moves all model parameters and buffers to the CPU.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cuda</span></code>([device])</p></td>
-<td><p>Moves all model parameters and buffers to the GPU.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">double</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">double</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">eval</span></code>()</p></td>
-<td><p>Sets the module in evaluation mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">extra_repr</span></code>()</p></td>
-<td><p>Set the extra representation of the module</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">float</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to float datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">half</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">half</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_state_dict</span></code>(state_dict[, strict])</p></td>
-<td><p>Copies parameters and buffers from <code class="xref py py-attr docutils literal notranslate"><span class="pre">state_dict</span></code> into this module and its descendants.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">modules</span></code>()</p></td>
-<td><p>Returns an iterator over all modules in the network.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_buffers</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_modules</span></code>([memo, prefix])</p></td>
-<td><p>Returns an iterator over all modules in the network, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_parameters</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parameters</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module parameters.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_backward_hook</span></code>(hook)</p></td>
-<td><p>Registers a backward hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_buffer</span></code>(name, tensor[, persistent])</p></td>
-<td><p>Adds a buffer to the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_pre_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward pre-hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_parameter</span></code>(name, param)</p></td>
-<td><p>Adds a parameter to the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">requires_grad_</span></code>([requires_grad])</p></td>
-<td><p>Change if autograd should record operations on parameters in this module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">same_padding</span></code>(d, k)</p></td>
-<td><p>Calculate the amount of padding.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">state_dict</span></code>([destination, prefix, keep_vars])</p></td>
-<td><p>Returns a dictionary containing a whole state of the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to</span></code>(*args, **kwargs)</p></td>
-<td><p>Moves and/or casts the parameters and buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">train</span></code>([mode])</p></td>
-<td><p>Sets the module in training mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">type</span></code>(dst_type)</p></td>
-<td><p>Casts all parameters and buffers to <code class="xref py py-attr docutils literal notranslate"><span class="pre">dst_type</span></code>.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">zero_grad</span></code>()</p></td>
-<td><p>Sets gradients of all model parameters to zero.</p></td>
-</tr>
-</tbody>
-</table>
-<table class="docutils align-default">
-<colgroup>
-<col style="width: 67%" />
-<col style="width: 33%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><strong>__call__</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-even"><td><p><strong>forward</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-odd"><td><p><strong>reset_parameters</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-even"><td><p><strong>share_memory</strong></p></td>
-<td></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="pysegcnn.core.layers.Conv2dSame.__init__">
-<code class="sig-name descname">__init__</code><span class="sig-paren">(</span><em class="sig-param"><span class="o">*</span><span class="n">args</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/layers.html#Conv2dSame.__init__"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.layers.Conv2dSame.__init__" title="Permalink to this definition">¶</a></dt>
-<dd><p>Initializes internal Module state, shared by both nn.Module and ScriptModule.</p>
</dd></dl>
<p class="rubric">Methods</p>
@@ -320,7 +220,7 @@ dimensions as the input.</p>
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#pysegcnn.core.layers.Conv2dSame.__init__" title="pysegcnn.core.layers.Conv2dSame.__init__"><code class="xref py py-obj docutils literal notranslate"><span class="pre">__init__</span></code></a>(*args, **kwargs)</p></td>
-<td><p>Initializes internal Module state, shared by both nn.Module and ScriptModule.</p></td>
+<td><p>Initialize.</p></td>
</tr>
<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
<td><p>Adds a child module to the current module.</p></td>
@@ -451,6 +351,15 @@ dimensions as the input.</p>
</div>
<footer>
+ <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
+
+ <a href="pysegcnn.core.layers.ConvBnReluMaxPool.html" class="btn btn-neutral float-right" title="pysegcnn.core.layers.ConvBnReluMaxPool" accesskey="n" rel="next">Next <span class="fa fa-arrow-circle-right"></span></a>
+
+
+ <a href="pysegcnn.core.layers.Block.html" class="btn btn-neutral float-left" title="pysegcnn.core.layers.Block" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left"></span> Previous</a>
+
+ </div>
+
<hr/>
diff --git a/docs/_build/html/source/generated/pysegcnn.core.layers.ConvBnReluMaxPool.html b/docs/_build/html/source/generated/pysegcnn.core.layers.ConvBnReluMaxPool.html
index 09c844e7b508d31111dc76a285696ca8092e3944..f037442d93026acc61b2876591af974b12ce80b9 100644
--- a/docs/_build/html/source/generated/pysegcnn.core.layers.ConvBnReluMaxPool.html
+++ b/docs/_build/html/source/generated/pysegcnn.core.layers.ConvBnReluMaxPool.html
@@ -35,7 +35,9 @@
<link rel="index" title="Index" href="../../genindex.html" />
- <link rel="search" title="Search" href="../../search.html" />
+ <link rel="search" title="Search" href="../../search.html" />
+ <link rel="next" title="pysegcnn.core.layers.ConvBnReluMaxUnpool" href="pysegcnn.core.layers.ConvBnReluMaxUnpool.html" />
+ <link rel="prev" title="pysegcnn.core.layers.Conv2dSame" href="pysegcnn.core.layers.Conv2dSame.html" />
</head>
<body class="wy-body-for-nav">
@@ -80,10 +82,26 @@
- <ul>
+ <ul class="current">
<li class="toctree-l1"><a class="reference internal" href="../requirements.html">Requirements</a></li>
<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../api.html">API Reference</a></li>
+<li class="toctree-l1 current"><a class="reference internal" href="../api.html">API Reference</a><ul class="current">
+<li class="toctree-l2"><a class="reference internal" href="../api.html#dataset">Dataset</a></li>
+<li class="toctree-l2 current"><a class="reference internal" href="../api.html#models">Models</a><ul class="current">
+<li class="toctree-l3 current"><a class="reference internal" href="../api.html#layers">Layers</a><ul class="current">
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.Block.html">pysegcnn.core.layers.Block</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.Conv2dSame.html">pysegcnn.core.layers.Conv2dSame</a></li>
+<li class="toctree-l4 current"><a class="current reference internal" href="#">pysegcnn.core.layers.ConvBnReluMaxPool</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.ConvBnReluMaxUnpool.html">pysegcnn.core.layers.ConvBnReluMaxUnpool</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.ConvBnReluUpsample.html">pysegcnn.core.layers.ConvBnReluUpsample</a></li>
+</ul>
+</li>
+<li class="toctree-l3"><a class="reference internal" href="../api.html#encoder-decoder-architechture">Encoder-Decoder architechture</a></li>
+<li class="toctree-l3"><a class="reference internal" href="../api.html#neural-networks">Neural Networks</a></li>
+</ul>
+</li>
+</ul>
+</li>
</ul>
@@ -130,6 +148,8 @@
<li><a href="../../index.html" class="icon icon-home"></a> »</li>
+ <li><a href="../api.html">API Reference</a> »</li>
+
<li>pysegcnn.core.layers.ConvBnReluMaxPool</li>
@@ -155,148 +175,30 @@
<dt id="pysegcnn.core.layers.ConvBnReluMaxPool">
<em class="property">class </em><code class="sig-prename descclassname">pysegcnn.core.layers.</code><code class="sig-name descname">ConvBnReluMaxPool</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">out_channels</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/layers.html#ConvBnReluMaxPool"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.layers.ConvBnReluMaxPool" title="Permalink to this definition">¶</a></dt>
<dd><p>Block of convolution, batchnorm, relu and 2x2 max pool.</p>
-<dl class="field-list simple">
+<dl class="py method">
+<dt id="pysegcnn.core.layers.ConvBnReluMaxPool.__init__">
+<code class="sig-name descname">__init__</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">out_channels</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/layers.html#ConvBnReluMaxPool.__init__"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.layers.ConvBnReluMaxPool.__init__" title="Permalink to this definition">¶</a></dt>
+<dd><p>Initialize.</p>
+<dl class="field-list">
<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>in_channels</strong> (<cite>int</cite>) – Number of input channels.</p></li>
-<li><p><strong>out_channels</strong> (<cite>int</cite>) – Number of output channels.</p></li>
-<li><p><strong>**kwargs</strong> (‘dict’ [<cite>str</cite>]) – Additional keyword arguments passed to
-<cite>pysegcnn.core.layers.Conv2dSame</cite>.</p></li>
-</ul>
+<dd class="field-odd"><dl>
+<dt><strong>in_channels</strong><span class="classifier"><cite>int</cite></span></dt><dd><p>Number of input channels.</p>
</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
+<dt><strong>out_channels</strong><span class="classifier"><cite>int</cite></span></dt><dd><p>Number of output channels.</p>
</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
+<dt><strong>**kwargs: `dict` [`str`]</strong></dt><dd><p>Additional arguments passed to
+<a class="reference internal" href="pysegcnn.core.layers.Conv2dSame.html#pysegcnn.core.layers.Conv2dSame" title="pysegcnn.core.layers.Conv2dSame"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.Conv2dSame</span></code></a>.</p>
+</dd>
+</dl>
+</dd>
+<dt class="field-even">Raises</dt>
+<dd class="field-even"><dl class="simple">
+<dt>TypeError</dt><dd><p>Raised if <code class="xref py py-meth docutils literal notranslate"><span class="pre">layers()</span></code> method does
+not return an instance of <code class="xref py py-class docutils literal notranslate"><span class="pre">torch.nn.Sequential</span></code>.</p>
+</dd>
+</dl>
</dd>
</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
-<td><p>Adds a child module to the current module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">apply</span></code>(fn)</p></td>
-<td><p>Applies <code class="docutils literal notranslate"><span class="pre">fn</span></code> recursively to every submodule (as returned by <code class="docutils literal notranslate"><span class="pre">.children()</span></code>) as well as self.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">bfloat16</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">bfloat16</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">buffers</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cpu</span></code>()</p></td>
-<td><p>Moves all model parameters and buffers to the CPU.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cuda</span></code>([device])</p></td>
-<td><p>Moves all model parameters and buffers to the GPU.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">double</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">double</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">downsample</span></code>(x)</p></td>
-<td><p>2x2 max pooling layer, <cite>torch.nn.functional.max_pool2d</cite>.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">eval</span></code>()</p></td>
-<td><p>Sets the module in evaluation mode.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">extra_repr</span></code>()</p></td>
-<td><p>Define optional extra information about this module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">float</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to float datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward</span></code>(x)</p></td>
-<td><p>Forward pass of an encoder block.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">half</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">half</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">layers</span></code>()</p></td>
-<td><p>Sequence of convolution, batchnorm and relu layers.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_state_dict</span></code>(state_dict[, strict])</p></td>
-<td><p>Copies parameters and buffers from <code class="xref py py-attr docutils literal notranslate"><span class="pre">state_dict</span></code> into this module and its descendants.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">modules</span></code>()</p></td>
-<td><p>Returns an iterator over all modules in the network.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_buffers</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_modules</span></code>([memo, prefix])</p></td>
-<td><p>Returns an iterator over all modules in the network, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_parameters</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parameters</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module parameters.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_backward_hook</span></code>(hook)</p></td>
-<td><p>Registers a backward hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_buffer</span></code>(name, tensor[, persistent])</p></td>
-<td><p>Adds a buffer to the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_pre_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward pre-hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_parameter</span></code>(name, param)</p></td>
-<td><p>Adds a parameter to the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">requires_grad_</span></code>([requires_grad])</p></td>
-<td><p>Change if autograd should record operations on parameters in this module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">state_dict</span></code>([destination, prefix, keep_vars])</p></td>
-<td><p>Returns a dictionary containing a whole state of the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to</span></code>(*args, **kwargs)</p></td>
-<td><p>Moves and/or casts the parameters and buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">train</span></code>([mode])</p></td>
-<td><p>Sets the module in training mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">type</span></code>(dst_type)</p></td>
-<td><p>Casts all parameters and buffers to <code class="xref py py-attr docutils literal notranslate"><span class="pre">dst_type</span></code>.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">zero_grad</span></code>()</p></td>
-<td><p>Sets gradients of all model parameters to zero.</p></td>
-</tr>
-</tbody>
-</table>
-<table class="docutils align-default">
-<colgroup>
-<col style="width: 62%" />
-<col style="width: 38%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><strong>__call__</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-even"><td><p><strong>share_memory</strong></p></td>
-<td></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="pysegcnn.core.layers.ConvBnReluMaxPool.__init__">
-<code class="sig-name descname">__init__</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">out_channels</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/layers.html#ConvBnReluMaxPool.__init__"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.layers.ConvBnReluMaxPool.__init__" title="Permalink to this definition">¶</a></dt>
-<dd><p>Initializes internal Module state, shared by both nn.Module and ScriptModule.</p>
</dd></dl>
<p class="rubric">Methods</p>
@@ -307,7 +209,7 @@
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#pysegcnn.core.layers.ConvBnReluMaxPool.__init__" title="pysegcnn.core.layers.ConvBnReluMaxPool.__init__"><code class="xref py py-obj docutils literal notranslate"><span class="pre">__init__</span></code></a>(in_channels, out_channels, **kwargs)</p></td>
-<td><p>Initializes internal Module state, shared by both nn.Module and ScriptModule.</p></td>
+<td><p>Initialize.</p></td>
</tr>
<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
<td><p>Adds a child module to the current module.</p></td>
@@ -334,7 +236,7 @@
<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">double</span></code> datatype.</p></td>
</tr>
<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">downsample</span></code>(x)</p></td>
-<td><p>2x2 max pooling layer, <cite>torch.nn.functional.max_pool2d</cite>.</p></td>
+<td><p>2x2 max pooling layer.</p></td>
</tr>
<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">eval</span></code>()</p></td>
<td><p>Sets the module in evaluation mode.</p></td>
@@ -438,6 +340,15 @@
</div>
<footer>
+ <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
+
+ <a href="pysegcnn.core.layers.ConvBnReluMaxUnpool.html" class="btn btn-neutral float-right" title="pysegcnn.core.layers.ConvBnReluMaxUnpool" accesskey="n" rel="next">Next <span class="fa fa-arrow-circle-right"></span></a>
+
+
+ <a href="pysegcnn.core.layers.Conv2dSame.html" class="btn btn-neutral float-left" title="pysegcnn.core.layers.Conv2dSame" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left"></span> Previous</a>
+
+ </div>
+
<hr/>
diff --git a/docs/_build/html/source/generated/pysegcnn.core.layers.ConvBnReluMaxUnpool.html b/docs/_build/html/source/generated/pysegcnn.core.layers.ConvBnReluMaxUnpool.html
index f694846b4861e74dbf7e2ae1eb15393389854a2b..3cbf3dc9ca97b3b3700b877bf31d5caeadbea5d2 100644
--- a/docs/_build/html/source/generated/pysegcnn.core.layers.ConvBnReluMaxUnpool.html
+++ b/docs/_build/html/source/generated/pysegcnn.core.layers.ConvBnReluMaxUnpool.html
@@ -35,7 +35,9 @@
<link rel="index" title="Index" href="../../genindex.html" />
- <link rel="search" title="Search" href="../../search.html" />
+ <link rel="search" title="Search" href="../../search.html" />
+ <link rel="next" title="pysegcnn.core.layers.ConvBnReluUpsample" href="pysegcnn.core.layers.ConvBnReluUpsample.html" />
+ <link rel="prev" title="pysegcnn.core.layers.ConvBnReluMaxPool" href="pysegcnn.core.layers.ConvBnReluMaxPool.html" />
</head>
<body class="wy-body-for-nav">
@@ -80,10 +82,26 @@
- <ul>
+ <ul class="current">
<li class="toctree-l1"><a class="reference internal" href="../requirements.html">Requirements</a></li>
<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../api.html">API Reference</a></li>
+<li class="toctree-l1 current"><a class="reference internal" href="../api.html">API Reference</a><ul class="current">
+<li class="toctree-l2"><a class="reference internal" href="../api.html#dataset">Dataset</a></li>
+<li class="toctree-l2 current"><a class="reference internal" href="../api.html#models">Models</a><ul class="current">
+<li class="toctree-l3 current"><a class="reference internal" href="../api.html#layers">Layers</a><ul class="current">
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.Block.html">pysegcnn.core.layers.Block</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.Conv2dSame.html">pysegcnn.core.layers.Conv2dSame</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.ConvBnReluMaxPool.html">pysegcnn.core.layers.ConvBnReluMaxPool</a></li>
+<li class="toctree-l4 current"><a class="current reference internal" href="#">pysegcnn.core.layers.ConvBnReluMaxUnpool</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.ConvBnReluUpsample.html">pysegcnn.core.layers.ConvBnReluUpsample</a></li>
+</ul>
+</li>
+<li class="toctree-l3"><a class="reference internal" href="../api.html#encoder-decoder-architechture">Encoder-Decoder architechture</a></li>
+<li class="toctree-l3"><a class="reference internal" href="../api.html#neural-networks">Neural Networks</a></li>
+</ul>
+</li>
+</ul>
+</li>
</ul>
@@ -130,6 +148,8 @@
<li><a href="../../index.html" class="icon icon-home"></a> »</li>
+ <li><a href="../api.html">API Reference</a> »</li>
+
<li>pysegcnn.core.layers.ConvBnReluMaxUnpool</li>
@@ -155,148 +175,30 @@
<dt id="pysegcnn.core.layers.ConvBnReluMaxUnpool">
<em class="property">class </em><code class="sig-prename descclassname">pysegcnn.core.layers.</code><code class="sig-name descname">ConvBnReluMaxUnpool</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">out_channels</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/layers.html#ConvBnReluMaxUnpool"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.layers.ConvBnReluMaxUnpool" title="Permalink to this definition">¶</a></dt>
<dd><p>Block of convolution, batchnorm, relu and 2x2 max unpool.</p>
-<dl class="field-list simple">
+<dl class="py method">
+<dt id="pysegcnn.core.layers.ConvBnReluMaxUnpool.__init__">
+<code class="sig-name descname">__init__</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">out_channels</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/layers.html#ConvBnReluMaxUnpool.__init__"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.layers.ConvBnReluMaxUnpool.__init__" title="Permalink to this definition">¶</a></dt>
+<dd><p>Initialize.</p>
+<dl class="field-list">
<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>in_channels</strong> (<cite>int</cite>) – Number of input channels.</p></li>
-<li><p><strong>out_channels</strong> (<cite>int</cite>) – Number of output channels</p></li>
-<li><p><strong>**kwargs</strong> (‘dict’ [<cite>str</cite>]) – Additional keyword arguments passed to
-<cite>pysegcnn.core.layers.Conv2dSame</cite>.</p></li>
-</ul>
+<dd class="field-odd"><dl>
+<dt><strong>in_channels</strong><span class="classifier"><cite>int</cite></span></dt><dd><p>Number of input channels.</p>
</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
+<dt><strong>out_channels</strong><span class="classifier"><cite>int</cite></span></dt><dd><p>Number of output channels.</p>
</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
+<dt><strong>**kwargs: `dict` [`str`]</strong></dt><dd><p>Additional arguments passed to
+<a class="reference internal" href="pysegcnn.core.layers.Conv2dSame.html#pysegcnn.core.layers.Conv2dSame" title="pysegcnn.core.layers.Conv2dSame"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.Conv2dSame</span></code></a>.</p>
+</dd>
+</dl>
+</dd>
+<dt class="field-even">Raises</dt>
+<dd class="field-even"><dl class="simple">
+<dt>TypeError</dt><dd><p>Raised if <code class="xref py py-meth docutils literal notranslate"><span class="pre">layers()</span></code> method does
+not return an instance of <code class="xref py py-class docutils literal notranslate"><span class="pre">torch.nn.Sequential</span></code>.</p>
+</dd>
+</dl>
</dd>
</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
-<td><p>Adds a child module to the current module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">apply</span></code>(fn)</p></td>
-<td><p>Applies <code class="docutils literal notranslate"><span class="pre">fn</span></code> recursively to every submodule (as returned by <code class="docutils literal notranslate"><span class="pre">.children()</span></code>) as well as self.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">bfloat16</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">bfloat16</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">buffers</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cpu</span></code>()</p></td>
-<td><p>Moves all model parameters and buffers to the CPU.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cuda</span></code>([device])</p></td>
-<td><p>Moves all model parameters and buffers to the GPU.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">double</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">double</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">eval</span></code>()</p></td>
-<td><p>Sets the module in evaluation mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">extra_repr</span></code>()</p></td>
-<td><p>Define optional extra information about this module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">float</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to float datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward</span></code>(x, feature, indices, skip)</p></td>
-<td><p>Forward pass of a decoder block.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">half</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">half</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">layers</span></code>()</p></td>
-<td><p>Sequence of convolution, batchnorm and relu layers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_state_dict</span></code>(state_dict[, strict])</p></td>
-<td><p>Copies parameters and buffers from <code class="xref py py-attr docutils literal notranslate"><span class="pre">state_dict</span></code> into this module and its descendants.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">modules</span></code>()</p></td>
-<td><p>Returns an iterator over all modules in the network.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_buffers</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_modules</span></code>([memo, prefix])</p></td>
-<td><p>Returns an iterator over all modules in the network, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_parameters</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parameters</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module parameters.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_backward_hook</span></code>(hook)</p></td>
-<td><p>Registers a backward hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_buffer</span></code>(name, tensor[, persistent])</p></td>
-<td><p>Adds a buffer to the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_pre_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward pre-hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_parameter</span></code>(name, param)</p></td>
-<td><p>Adds a parameter to the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">requires_grad_</span></code>([requires_grad])</p></td>
-<td><p>Change if autograd should record operations on parameters in this module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">state_dict</span></code>([destination, prefix, keep_vars])</p></td>
-<td><p>Returns a dictionary containing a whole state of the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to</span></code>(*args, **kwargs)</p></td>
-<td><p>Moves and/or casts the parameters and buffers.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">train</span></code>([mode])</p></td>
-<td><p>Sets the module in training mode.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">type</span></code>(dst_type)</p></td>
-<td><p>Casts all parameters and buffers to <code class="xref py py-attr docutils literal notranslate"><span class="pre">dst_type</span></code>.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">upsample</span></code>(x, feature, indices)</p></td>
-<td><p>2x2 max unpooling layer.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">zero_grad</span></code>()</p></td>
-<td><p>Sets gradients of all model parameters to zero.</p></td>
-</tr>
-</tbody>
-</table>
-<table class="docutils align-default">
-<colgroup>
-<col style="width: 62%" />
-<col style="width: 38%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><strong>__call__</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-even"><td><p><strong>share_memory</strong></p></td>
-<td></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="pysegcnn.core.layers.ConvBnReluMaxUnpool.__init__">
-<code class="sig-name descname">__init__</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">out_channels</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/layers.html#ConvBnReluMaxUnpool.__init__"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.layers.ConvBnReluMaxUnpool.__init__" title="Permalink to this definition">¶</a></dt>
-<dd><p>Initializes internal Module state, shared by both nn.Module and ScriptModule.</p>
</dd></dl>
<p class="rubric">Methods</p>
@@ -307,7 +209,7 @@
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#pysegcnn.core.layers.ConvBnReluMaxUnpool.__init__" title="pysegcnn.core.layers.ConvBnReluMaxUnpool.__init__"><code class="xref py py-obj docutils literal notranslate"><span class="pre">__init__</span></code></a>(in_channels, out_channels, **kwargs)</p></td>
-<td><p>Initializes internal Module state, shared by both nn.Module and ScriptModule.</p></td>
+<td><p>Initialize.</p></td>
</tr>
<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
<td><p>Adds a child module to the current module.</p></td>
@@ -438,6 +340,15 @@
</div>
<footer>
+ <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
+
+ <a href="pysegcnn.core.layers.ConvBnReluUpsample.html" class="btn btn-neutral float-right" title="pysegcnn.core.layers.ConvBnReluUpsample" accesskey="n" rel="next">Next <span class="fa fa-arrow-circle-right"></span></a>
+
+
+ <a href="pysegcnn.core.layers.ConvBnReluMaxPool.html" class="btn btn-neutral float-left" title="pysegcnn.core.layers.ConvBnReluMaxPool" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left"></span> Previous</a>
+
+ </div>
+
<hr/>
diff --git a/docs/_build/html/source/generated/pysegcnn.core.layers.ConvBnReluUpsample.html b/docs/_build/html/source/generated/pysegcnn.core.layers.ConvBnReluUpsample.html
index f5ebbff95d8836755c94d745ec89d9f14911f37a..c504ad0ca50329c93cc7f67fb37d631317c30688 100644
--- a/docs/_build/html/source/generated/pysegcnn.core.layers.ConvBnReluUpsample.html
+++ b/docs/_build/html/source/generated/pysegcnn.core.layers.ConvBnReluUpsample.html
@@ -35,7 +35,9 @@
<link rel="index" title="Index" href="../../genindex.html" />
- <link rel="search" title="Search" href="../../search.html" />
+ <link rel="search" title="Search" href="../../search.html" />
+ <link rel="next" title="pysegcnn.core.layers.EncoderBlock" href="pysegcnn.core.layers.EncoderBlock.html" />
+ <link rel="prev" title="pysegcnn.core.layers.ConvBnReluMaxUnpool" href="pysegcnn.core.layers.ConvBnReluMaxUnpool.html" />
</head>
<body class="wy-body-for-nav">
@@ -80,10 +82,26 @@
- <ul>
+ <ul class="current">
<li class="toctree-l1"><a class="reference internal" href="../requirements.html">Requirements</a></li>
<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../api.html">API Reference</a></li>
+<li class="toctree-l1 current"><a class="reference internal" href="../api.html">API Reference</a><ul class="current">
+<li class="toctree-l2"><a class="reference internal" href="../api.html#dataset">Dataset</a></li>
+<li class="toctree-l2 current"><a class="reference internal" href="../api.html#models">Models</a><ul class="current">
+<li class="toctree-l3 current"><a class="reference internal" href="../api.html#layers">Layers</a><ul class="current">
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.Block.html">pysegcnn.core.layers.Block</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.Conv2dSame.html">pysegcnn.core.layers.Conv2dSame</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.ConvBnReluMaxPool.html">pysegcnn.core.layers.ConvBnReluMaxPool</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.ConvBnReluMaxUnpool.html">pysegcnn.core.layers.ConvBnReluMaxUnpool</a></li>
+<li class="toctree-l4 current"><a class="current reference internal" href="#">pysegcnn.core.layers.ConvBnReluUpsample</a></li>
+</ul>
+</li>
+<li class="toctree-l3"><a class="reference internal" href="../api.html#encoder-decoder-architechture">Encoder-Decoder architechture</a></li>
+<li class="toctree-l3"><a class="reference internal" href="../api.html#neural-networks">Neural Networks</a></li>
+</ul>
+</li>
+</ul>
+</li>
</ul>
@@ -130,6 +148,8 @@
<li><a href="../../index.html" class="icon icon-home"></a> »</li>
+ <li><a href="../api.html">API Reference</a> »</li>
+
<li>pysegcnn.core.layers.ConvBnReluUpsample</li>
@@ -154,148 +174,31 @@
<dl class="py class">
<dt id="pysegcnn.core.layers.ConvBnReluUpsample">
<em class="property">class </em><code class="sig-prename descclassname">pysegcnn.core.layers.</code><code class="sig-name descname">ConvBnReluUpsample</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">out_channels</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/layers.html#ConvBnReluUpsample"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.layers.ConvBnReluUpsample" title="Permalink to this definition">¶</a></dt>
-<dd><p>Block of convolution, batchnorm, relu and nearest neighbor upsampling.</p>
-<dl class="field-list simple">
+<dd><p>Block of convolution, batchnorm, relu and nearest neighbor upsample.</p>
+<dl class="py method">
+<dt id="pysegcnn.core.layers.ConvBnReluUpsample.__init__">
+<code class="sig-name descname">__init__</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">out_channels</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/layers.html#ConvBnReluUpsample.__init__"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.layers.ConvBnReluUpsample.__init__" title="Permalink to this definition">¶</a></dt>
+<dd><p>Initialize.</p>
+<dl class="field-list">
<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>in_channels</strong> (<cite>int</cite>) – Number of input channels.</p></li>
-<li><p><strong>out_channels</strong> (<cite>int</cite>) – Number of output channels</p></li>
-<li><p><strong>**kwargs</strong> (‘dict’ [<cite>str</cite>]) – Additional arguments passed to <cite>pysegcnn.core.layers.Conv2dSame</cite>.</p></li>
-</ul>
+<dd class="field-odd"><dl>
+<dt><strong>in_channels</strong><span class="classifier"><cite>int</cite></span></dt><dd><p>Number of input channels.</p>
</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
+<dt><strong>out_channels</strong><span class="classifier"><cite>int</cite></span></dt><dd><p>Number of output channels.</p>
</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
+<dt><strong>**kwargs: `dict` [`str`]</strong></dt><dd><p>Additional arguments passed to
+<a class="reference internal" href="pysegcnn.core.layers.Conv2dSame.html#pysegcnn.core.layers.Conv2dSame" title="pysegcnn.core.layers.Conv2dSame"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.Conv2dSame</span></code></a>.</p>
+</dd>
+</dl>
+</dd>
+<dt class="field-even">Raises</dt>
+<dd class="field-even"><dl class="simple">
+<dt>TypeError</dt><dd><p>Raised if <code class="xref py py-meth docutils literal notranslate"><span class="pre">layers()</span></code> method does
+not return an instance of <code class="xref py py-class docutils literal notranslate"><span class="pre">torch.nn.Sequential</span></code>.</p>
+</dd>
+</dl>
</dd>
</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
-<td><p>Adds a child module to the current module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">apply</span></code>(fn)</p></td>
-<td><p>Applies <code class="docutils literal notranslate"><span class="pre">fn</span></code> recursively to every submodule (as returned by <code class="docutils literal notranslate"><span class="pre">.children()</span></code>) as well as self.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">bfloat16</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">bfloat16</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">buffers</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cpu</span></code>()</p></td>
-<td><p>Moves all model parameters and buffers to the CPU.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cuda</span></code>([device])</p></td>
-<td><p>Moves all model parameters and buffers to the GPU.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">double</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">double</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">eval</span></code>()</p></td>
-<td><p>Sets the module in evaluation mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">extra_repr</span></code>()</p></td>
-<td><p>Define optional extra information about this module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">float</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to float datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward</span></code>(x, feature, indices, skip)</p></td>
-<td><p>Forward pass of a decoder block.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">half</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">half</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">layers</span></code>()</p></td>
-<td><p>Sequence of convolution, batchnorm and relu layers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_state_dict</span></code>(state_dict[, strict])</p></td>
-<td><p>Copies parameters and buffers from <code class="xref py py-attr docutils literal notranslate"><span class="pre">state_dict</span></code> into this module and its descendants.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">modules</span></code>()</p></td>
-<td><p>Returns an iterator over all modules in the network.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_buffers</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_modules</span></code>([memo, prefix])</p></td>
-<td><p>Returns an iterator over all modules in the network, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_parameters</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parameters</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module parameters.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_backward_hook</span></code>(hook)</p></td>
-<td><p>Registers a backward hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_buffer</span></code>(name, tensor[, persistent])</p></td>
-<td><p>Adds a buffer to the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_pre_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward pre-hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_parameter</span></code>(name, param)</p></td>
-<td><p>Adds a parameter to the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">requires_grad_</span></code>([requires_grad])</p></td>
-<td><p>Change if autograd should record operations on parameters in this module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">state_dict</span></code>([destination, prefix, keep_vars])</p></td>
-<td><p>Returns a dictionary containing a whole state of the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to</span></code>(*args, **kwargs)</p></td>
-<td><p>Moves and/or casts the parameters and buffers.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">train</span></code>([mode])</p></td>
-<td><p>Sets the module in training mode.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">type</span></code>(dst_type)</p></td>
-<td><p>Casts all parameters and buffers to <code class="xref py py-attr docutils literal notranslate"><span class="pre">dst_type</span></code>.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">upsample</span></code>(x, feature[, indices])</p></td>
-<td><p>Nearest neighbor upsampling.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">zero_grad</span></code>()</p></td>
-<td><p>Sets gradients of all model parameters to zero.</p></td>
-</tr>
-</tbody>
-</table>
-<table class="docutils align-default">
-<colgroup>
-<col style="width: 62%" />
-<col style="width: 38%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><strong>__call__</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-even"><td><p><strong>share_memory</strong></p></td>
-<td></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="pysegcnn.core.layers.ConvBnReluUpsample.__init__">
-<code class="sig-name descname">__init__</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">out_channels</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/layers.html#ConvBnReluUpsample.__init__"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.layers.ConvBnReluUpsample.__init__" title="Permalink to this definition">¶</a></dt>
-<dd><p>Initializes internal Module state, shared by both nn.Module and ScriptModule.</p>
</dd></dl>
<p class="rubric">Methods</p>
@@ -306,7 +209,7 @@
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#pysegcnn.core.layers.ConvBnReluUpsample.__init__" title="pysegcnn.core.layers.ConvBnReluUpsample.__init__"><code class="xref py py-obj docutils literal notranslate"><span class="pre">__init__</span></code></a>(in_channels, out_channels, **kwargs)</p></td>
-<td><p>Initializes internal Module state, shared by both nn.Module and ScriptModule.</p></td>
+<td><p>Initialize.</p></td>
</tr>
<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
<td><p>Adds a child module to the current module.</p></td>
@@ -437,6 +340,15 @@
</div>
<footer>
+ <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
+
+ <a href="pysegcnn.core.layers.EncoderBlock.html" class="btn btn-neutral float-right" title="pysegcnn.core.layers.EncoderBlock" accesskey="n" rel="next">Next <span class="fa fa-arrow-circle-right"></span></a>
+
+
+ <a href="pysegcnn.core.layers.ConvBnReluMaxUnpool.html" class="btn btn-neutral float-left" title="pysegcnn.core.layers.ConvBnReluMaxUnpool" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left"></span> Previous</a>
+
+ </div>
+
<hr/>
diff --git a/docs/_build/html/source/generated/pysegcnn.core.layers.Decoder.html b/docs/_build/html/source/generated/pysegcnn.core.layers.Decoder.html
index 286b5ca0d9a0c8c74b1253e1459deb7168a39357..761c49f489f68b8851138a325bc468bbf796e155 100644
--- a/docs/_build/html/source/generated/pysegcnn.core.layers.Decoder.html
+++ b/docs/_build/html/source/generated/pysegcnn.core.layers.Decoder.html
@@ -35,7 +35,9 @@
<link rel="index" title="Index" href="../../genindex.html" />
- <link rel="search" title="Search" href="../../search.html" />
+ <link rel="search" title="Search" href="../../search.html" />
+ <link rel="next" title="pysegcnn.core.models.Network" href="pysegcnn.core.models.Network.html" />
+ <link rel="prev" title="pysegcnn.core.layers.Encoder" href="pysegcnn.core.layers.Encoder.html" />
</head>
<body class="wy-body-for-nav">
@@ -80,10 +82,25 @@
- <ul>
+ <ul class="current">
<li class="toctree-l1"><a class="reference internal" href="../requirements.html">Requirements</a></li>
<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../api.html">API Reference</a></li>
+<li class="toctree-l1 current"><a class="reference internal" href="../api.html">API Reference</a><ul class="current">
+<li class="toctree-l2"><a class="reference internal" href="../api.html#dataset">Dataset</a></li>
+<li class="toctree-l2 current"><a class="reference internal" href="../api.html#models">Models</a><ul class="current">
+<li class="toctree-l3"><a class="reference internal" href="../api.html#layers">Layers</a></li>
+<li class="toctree-l3 current"><a class="reference internal" href="../api.html#encoder-decoder-architechture">Encoder-Decoder architechture</a><ul class="current">
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.EncoderBlock.html">pysegcnn.core.layers.EncoderBlock</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.DecoderBlock.html">pysegcnn.core.layers.DecoderBlock</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.Encoder.html">pysegcnn.core.layers.Encoder</a></li>
+<li class="toctree-l4 current"><a class="current reference internal" href="#">pysegcnn.core.layers.Decoder</a></li>
+</ul>
+</li>
+<li class="toctree-l3"><a class="reference internal" href="../api.html#neural-networks">Neural Networks</a></li>
+</ul>
+</li>
+</ul>
+</li>
</ul>
@@ -130,6 +147,8 @@
<li><a href="../../index.html" class="icon icon-home"></a> »</li>
+ <li><a href="../api.html">API Reference</a> »</li>
+
<li>pysegcnn.core.layers.Decoder</li>
@@ -156,149 +175,49 @@
<em class="property">class </em><code class="sig-prename descclassname">pysegcnn.core.layers.</code><code class="sig-name descname">Decoder</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">filters</span></em>, <em class="sig-param"><span class="n">block</span></em>, <em class="sig-param"><span class="n">skip</span><span class="o">=</span><span class="default_value">True</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/layers.html#Decoder"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.layers.Decoder" title="Permalink to this definition">¶</a></dt>
<dd><p>Generic convolutional decoder.</p>
<p>When instanciating an encoder-decoder architechure, <code class="docutils literal notranslate"><span class="pre">filters</span></code> should be
-the same for <cite>pysegcnn.core.layers.Encoder</cite> and
-<cite>pysegcnn.core.layers.Decoder</cite>.</p>
-<p>See <cite>pysegcnn.core.models.UNet</cite> for an example implementation.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>filters</strong> (<cite>list</cite> [<cite>int</cite>]) – List of input channels to each convolutional block. The length of
-<code class="docutils literal notranslate"><span class="pre">filters</span></code> determines the depth of the decoder. The first element of
-<code class="docutils literal notranslate"><span class="pre">filters</span></code> has to be the number of channels of the input images.</p></li>
-<li><p><strong>block</strong> (<cite>pysegcnn.core.layers.DecoderBlock</cite>) – The convolutional block defining a layer in the decoder.
-A subclass of <cite>pysegcnn.core.layers.DecoderBlock</cite>, e.g.
-<cite>pysegcnn.core.layers.ConvBnReluMaxUnpool</cite>.</p></li>
-<li><p><strong>skip</strong> (<cite>bool</cite>) – Whether to apply skip connections from the encoder to the decoder.</p></li>
-<li><p><strong>**kwargs</strong> (‘dict’ [<cite>str</cite>]) – Additional arguments passed to <cite>pysegcnn.core.layers.Conv2dSame</cite>.</p></li>
-</ul>
+the same for <a class="reference internal" href="pysegcnn.core.layers.Encoder.html#pysegcnn.core.layers.Encoder" title="pysegcnn.core.layers.Encoder"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.Encoder</span></code></a> and
+<a class="reference internal" href="#pysegcnn.core.layers.Decoder" title="pysegcnn.core.layers.Decoder"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.Decoder</span></code></a>.</p>
+<p>See <a class="reference internal" href="pysegcnn.core.models.UNet.html#pysegcnn.core.models.UNet" title="pysegcnn.core.models.UNet"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.models.UNet</span></code></a> for an example implementation.</p>
+<dl class="field-list">
+<dt class="field-odd">Attributes</dt>
+<dd class="field-odd"><dl>
+<dt><strong>features</strong><span class="classifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">numpy.ndarray</span></code></span></dt><dd><p>Input channels to each convolutional block, i.e. <code class="docutils literal notranslate"><span class="pre">filters</span></code>.</p>
+</dd>
+<dt><strong>block</strong><span class="classifier"><a class="reference internal" href="pysegcnn.core.layers.EncoderBlock.html#pysegcnn.core.layers.EncoderBlock" title="pysegcnn.core.layers.EncoderBlock"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.EncoderBlock</span></code></a></span></dt><dd><p>The convolutional block defining a layer in the decoder.
+A subclass of <a class="reference internal" href="pysegcnn.core.layers.EncoderBlock.html#pysegcnn.core.layers.EncoderBlock" title="pysegcnn.core.layers.EncoderBlock"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.EncoderBlock</span></code></a>, e.g.
+<a class="reference internal" href="pysegcnn.core.layers.ConvBnReluMaxPool.html#pysegcnn.core.layers.ConvBnReluMaxPool" title="pysegcnn.core.layers.ConvBnReluMaxPool"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.ConvBnReluMaxPool</span></code></a>.</p>
</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
+<dt><strong>skip</strong><span class="classifier"><cite>bool</cite></span></dt><dd><p>Whether to apply skip connections from the encoder to the decoder.</p>
</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
+<dt><strong>layers</strong><span class="classifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">torch.nn.ModuleList</span></code></span></dt><dd><p>List of blocks in the decoder.</p>
+</dd>
+</dl>
</dd>
</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
-<td><p>Adds a child module to the current module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">apply</span></code>(fn)</p></td>
-<td><p>Applies <code class="docutils literal notranslate"><span class="pre">fn</span></code> recursively to every submodule (as returned by <code class="docutils literal notranslate"><span class="pre">.children()</span></code>) as well as self.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">bfloat16</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">bfloat16</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">buffers</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cpu</span></code>()</p></td>
-<td><p>Moves all model parameters and buffers to the CPU.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cuda</span></code>([device])</p></td>
-<td><p>Moves all model parameters and buffers to the GPU.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">double</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">double</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">eval</span></code>()</p></td>
-<td><p>Sets the module in evaluation mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">extra_repr</span></code>()</p></td>
-<td><p>Set the extra representation of the module</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">float</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to float datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward</span></code>(x, enc_cache)</p></td>
-<td><p>Forward pass of the decoder.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">half</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">half</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_state_dict</span></code>(state_dict[, strict])</p></td>
-<td><p>Copies parameters and buffers from <code class="xref py py-attr docutils literal notranslate"><span class="pre">state_dict</span></code> into this module and its descendants.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">modules</span></code>()</p></td>
-<td><p>Returns an iterator over all modules in the network.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_buffers</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_modules</span></code>([memo, prefix])</p></td>
-<td><p>Returns an iterator over all modules in the network, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_parameters</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parameters</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module parameters.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_backward_hook</span></code>(hook)</p></td>
-<td><p>Registers a backward hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_buffer</span></code>(name, tensor[, persistent])</p></td>
-<td><p>Adds a buffer to the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_pre_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward pre-hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_parameter</span></code>(name, param)</p></td>
-<td><p>Adds a parameter to the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">requires_grad_</span></code>([requires_grad])</p></td>
-<td><p>Change if autograd should record operations on parameters in this module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">state_dict</span></code>([destination, prefix, keep_vars])</p></td>
-<td><p>Returns a dictionary containing a whole state of the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to</span></code>(*args, **kwargs)</p></td>
-<td><p>Moves and/or casts the parameters and buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">train</span></code>([mode])</p></td>
-<td><p>Sets the module in training mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">type</span></code>(dst_type)</p></td>
-<td><p>Casts all parameters and buffers to <code class="xref py py-attr docutils literal notranslate"><span class="pre">dst_type</span></code>.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">zero_grad</span></code>()</p></td>
-<td><p>Sets gradients of all model parameters to zero.</p></td>
-</tr>
-</tbody>
-</table>
-<table class="docutils align-default">
-<colgroup>
-<col style="width: 62%" />
-<col style="width: 38%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><strong>__call__</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-even"><td><p><strong>share_memory</strong></p></td>
-<td></td>
-</tr>
-</tbody>
-</table>
<dl class="py method">
<dt id="pysegcnn.core.layers.Decoder.__init__">
<code class="sig-name descname">__init__</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">filters</span></em>, <em class="sig-param"><span class="n">block</span></em>, <em class="sig-param"><span class="n">skip</span><span class="o">=</span><span class="default_value">True</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/layers.html#Decoder.__init__"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.layers.Decoder.__init__" title="Permalink to this definition">¶</a></dt>
-<dd><p>Initializes internal Module state, shared by both nn.Module and ScriptModule.</p>
+<dd><p>Initialize.</p>
+<dl class="field-list">
+<dt class="field-odd">Parameters</dt>
+<dd class="field-odd"><dl>
+<dt><strong>filters</strong><span class="classifier"><cite>list</cite> [<cite>int</cite>]</span></dt><dd><p>List of input channels to each convolutional block. The length of
+<code class="docutils literal notranslate"><span class="pre">filters</span></code> determines the depth of the decoder. The first element
+of <code class="docutils literal notranslate"><span class="pre">filters</span></code> has to be the number of channels of the input
+images.</p>
+</dd>
+<dt><strong>block</strong><span class="classifier"><a class="reference internal" href="pysegcnn.core.layers.DecoderBlock.html#pysegcnn.core.layers.DecoderBlock" title="pysegcnn.core.layers.DecoderBlock"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.DecoderBlock</span></code></a></span></dt><dd><p>The convolutional block defining a layer in the decoder.
+A subclass of <a class="reference internal" href="pysegcnn.core.layers.DecoderBlock.html#pysegcnn.core.layers.DecoderBlock" title="pysegcnn.core.layers.DecoderBlock"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.DecoderBlock</span></code></a>, e.g.
+<a class="reference internal" href="pysegcnn.core.layers.ConvBnReluMaxUnpool.html#pysegcnn.core.layers.ConvBnReluMaxUnpool" title="pysegcnn.core.layers.ConvBnReluMaxUnpool"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.ConvBnReluMaxUnpool</span></code></a>.</p>
+</dd>
+<dt><strong>skip</strong><span class="classifier"><cite>bool</cite>, optional</span></dt><dd><p>Whether to apply skip connections from the encoder to the decoder.</p>
+</dd>
+<dt><strong>**kwargs: `dict` [`str`]</strong></dt><dd><p>Additional arguments passed to
+<a class="reference internal" href="pysegcnn.core.layers.Conv2dSame.html#pysegcnn.core.layers.Conv2dSame" title="pysegcnn.core.layers.Conv2dSame"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.Conv2dSame</span></code></a>.</p>
+</dd>
+</dl>
+</dd>
+</dl>
</dd></dl>
<p class="rubric">Methods</p>
@@ -309,7 +228,7 @@ A subclass of <cite>pysegcnn.core.layers.DecoderBlock</cite>, e.g.
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#pysegcnn.core.layers.Decoder.__init__" title="pysegcnn.core.layers.Decoder.__init__"><code class="xref py py-obj docutils literal notranslate"><span class="pre">__init__</span></code></a>(filters, block[, skip])</p></td>
-<td><p>Initializes internal Module state, shared by both nn.Module and ScriptModule.</p></td>
+<td><p>Initialize.</p></td>
</tr>
<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
<td><p>Adds a child module to the current module.</p></td>
@@ -434,6 +353,15 @@ A subclass of <cite>pysegcnn.core.layers.DecoderBlock</cite>, e.g.
</div>
<footer>
+ <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
+
+ <a href="pysegcnn.core.models.Network.html" class="btn btn-neutral float-right" title="pysegcnn.core.models.Network" accesskey="n" rel="next">Next <span class="fa fa-arrow-circle-right"></span></a>
+
+
+ <a href="pysegcnn.core.layers.Encoder.html" class="btn btn-neutral float-left" title="pysegcnn.core.layers.Encoder" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left"></span> Previous</a>
+
+ </div>
+
<hr/>
diff --git a/docs/_build/html/source/generated/pysegcnn.core.layers.DecoderBlock.html b/docs/_build/html/source/generated/pysegcnn.core.layers.DecoderBlock.html
index fa52e72eda7a7c3c46a1c65fc80776da7ca8b856..830aa5d909d6cc11ac9149025e5c0fe5750c5a6b 100644
--- a/docs/_build/html/source/generated/pysegcnn.core.layers.DecoderBlock.html
+++ b/docs/_build/html/source/generated/pysegcnn.core.layers.DecoderBlock.html
@@ -35,7 +35,9 @@
<link rel="index" title="Index" href="../../genindex.html" />
- <link rel="search" title="Search" href="../../search.html" />
+ <link rel="search" title="Search" href="../../search.html" />
+ <link rel="next" title="pysegcnn.core.layers.Encoder" href="pysegcnn.core.layers.Encoder.html" />
+ <link rel="prev" title="pysegcnn.core.layers.EncoderBlock" href="pysegcnn.core.layers.EncoderBlock.html" />
</head>
<body class="wy-body-for-nav">
@@ -80,10 +82,25 @@
- <ul>
+ <ul class="current">
<li class="toctree-l1"><a class="reference internal" href="../requirements.html">Requirements</a></li>
<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../api.html">API Reference</a></li>
+<li class="toctree-l1 current"><a class="reference internal" href="../api.html">API Reference</a><ul class="current">
+<li class="toctree-l2"><a class="reference internal" href="../api.html#dataset">Dataset</a></li>
+<li class="toctree-l2 current"><a class="reference internal" href="../api.html#models">Models</a><ul class="current">
+<li class="toctree-l3"><a class="reference internal" href="../api.html#layers">Layers</a></li>
+<li class="toctree-l3 current"><a class="reference internal" href="../api.html#encoder-decoder-architechture">Encoder-Decoder architechture</a><ul class="current">
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.EncoderBlock.html">pysegcnn.core.layers.EncoderBlock</a></li>
+<li class="toctree-l4 current"><a class="current reference internal" href="#">pysegcnn.core.layers.DecoderBlock</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.Encoder.html">pysegcnn.core.layers.Encoder</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.Decoder.html">pysegcnn.core.layers.Decoder</a></li>
+</ul>
+</li>
+<li class="toctree-l3"><a class="reference internal" href="../api.html#neural-networks">Neural Networks</a></li>
+</ul>
+</li>
+</ul>
+</li>
</ul>
@@ -130,6 +147,8 @@
<li><a href="../../index.html" class="icon icon-home"></a> »</li>
+ <li><a href="../api.html">API Reference</a> »</li>
+
<li>pysegcnn.core.layers.DecoderBlock</li>
@@ -155,147 +174,30 @@
<dt id="pysegcnn.core.layers.DecoderBlock">
<em class="property">class </em><code class="sig-prename descclassname">pysegcnn.core.layers.</code><code class="sig-name descname">DecoderBlock</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">out_channels</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/layers.html#DecoderBlock"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.layers.DecoderBlock" title="Permalink to this definition">¶</a></dt>
<dd><p>Block of a convolutional decoder.</p>
-<dl class="field-list simple">
+<dl class="py method">
+<dt id="pysegcnn.core.layers.DecoderBlock.__init__">
+<code class="sig-name descname">__init__</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">out_channels</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/layers.html#DecoderBlock.__init__"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.layers.DecoderBlock.__init__" title="Permalink to this definition">¶</a></dt>
+<dd><p>Initialize.</p>
+<dl class="field-list">
<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>in_channels</strong> (<cite>int</cite>) – Number of input channels.</p></li>
-<li><p><strong>out_channels</strong> (<cite>int</cite>) – Number of output channels.</p></li>
-<li><p><strong>**kwargs</strong> (‘dict’ [<cite>str</cite>]) – Additional arguments passed to <cite>pysegcnn.core.layers.Conv2dSame</cite>.</p></li>
-</ul>
+<dd class="field-odd"><dl>
+<dt><strong>in_channels</strong><span class="classifier"><cite>int</cite></span></dt><dd><p>Number of input channels.</p>
</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
+<dt><strong>out_channels</strong><span class="classifier"><cite>int</cite></span></dt><dd><p>Number of output channels.</p>
</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
+<dt><strong>**kwargs: `dict` [`str`]</strong></dt><dd><p>Additional arguments passed to
+<a class="reference internal" href="pysegcnn.core.layers.Conv2dSame.html#pysegcnn.core.layers.Conv2dSame" title="pysegcnn.core.layers.Conv2dSame"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.Conv2dSame</span></code></a>.</p>
+</dd>
+</dl>
+</dd>
+<dt class="field-even">Raises</dt>
+<dd class="field-even"><dl class="simple">
+<dt>TypeError</dt><dd><p>Raised if <code class="xref py py-meth docutils literal notranslate"><span class="pre">layers()</span></code> method does
+not return an instance of <code class="xref py py-class docutils literal notranslate"><span class="pre">torch.nn.Sequential</span></code>.</p>
+</dd>
+</dl>
</dd>
</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
-<td><p>Adds a child module to the current module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">apply</span></code>(fn)</p></td>
-<td><p>Applies <code class="docutils literal notranslate"><span class="pre">fn</span></code> recursively to every submodule (as returned by <code class="docutils literal notranslate"><span class="pre">.children()</span></code>) as well as self.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">bfloat16</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">bfloat16</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">buffers</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cpu</span></code>()</p></td>
-<td><p>Moves all model parameters and buffers to the CPU.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cuda</span></code>([device])</p></td>
-<td><p>Moves all model parameters and buffers to the GPU.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">double</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">double</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">eval</span></code>()</p></td>
-<td><p>Sets the module in evaluation mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">extra_repr</span></code>()</p></td>
-<td><p>Set the extra representation of the module</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">float</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to float datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward</span></code>(x, feature, indices, skip)</p></td>
-<td><p>Forward pass of a decoder block.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">half</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">half</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">layers</span></code>()</p></td>
-<td><p>Define the layers of the block.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_state_dict</span></code>(state_dict[, strict])</p></td>
-<td><p>Copies parameters and buffers from <code class="xref py py-attr docutils literal notranslate"><span class="pre">state_dict</span></code> into this module and its descendants.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">modules</span></code>()</p></td>
-<td><p>Returns an iterator over all modules in the network.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_buffers</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_modules</span></code>([memo, prefix])</p></td>
-<td><p>Returns an iterator over all modules in the network, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_parameters</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parameters</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module parameters.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_backward_hook</span></code>(hook)</p></td>
-<td><p>Registers a backward hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_buffer</span></code>(name, tensor[, persistent])</p></td>
-<td><p>Adds a buffer to the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_pre_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward pre-hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_parameter</span></code>(name, param)</p></td>
-<td><p>Adds a parameter to the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">requires_grad_</span></code>([requires_grad])</p></td>
-<td><p>Change if autograd should record operations on parameters in this module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">state_dict</span></code>([destination, prefix, keep_vars])</p></td>
-<td><p>Returns a dictionary containing a whole state of the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to</span></code>(*args, **kwargs)</p></td>
-<td><p>Moves and/or casts the parameters and buffers.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">train</span></code>([mode])</p></td>
-<td><p>Sets the module in training mode.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">type</span></code>(dst_type)</p></td>
-<td><p>Casts all parameters and buffers to <code class="xref py py-attr docutils literal notranslate"><span class="pre">dst_type</span></code>.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">upsample</span></code>(x, feature, indices)</p></td>
-<td><p>Define the upsampling method.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">zero_grad</span></code>()</p></td>
-<td><p>Sets gradients of all model parameters to zero.</p></td>
-</tr>
-</tbody>
-</table>
-<table class="docutils align-default">
-<colgroup>
-<col style="width: 62%" />
-<col style="width: 38%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><strong>__call__</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-even"><td><p><strong>share_memory</strong></p></td>
-<td></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="pysegcnn.core.layers.DecoderBlock.__init__">
-<code class="sig-name descname">__init__</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">out_channels</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/layers.html#DecoderBlock.__init__"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.layers.DecoderBlock.__init__" title="Permalink to this definition">¶</a></dt>
-<dd><p>Initializes internal Module state, shared by both nn.Module and ScriptModule.</p>
</dd></dl>
<p class="rubric">Methods</p>
@@ -306,7 +208,7 @@
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#pysegcnn.core.layers.DecoderBlock.__init__" title="pysegcnn.core.layers.DecoderBlock.__init__"><code class="xref py py-obj docutils literal notranslate"><span class="pre">__init__</span></code></a>(in_channels, out_channels, **kwargs)</p></td>
-<td><p>Initializes internal Module state, shared by both nn.Module and ScriptModule.</p></td>
+<td><p>Initialize.</p></td>
</tr>
<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
<td><p>Adds a child module to the current module.</p></td>
@@ -437,6 +339,15 @@
</div>
<footer>
+ <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
+
+ <a href="pysegcnn.core.layers.Encoder.html" class="btn btn-neutral float-right" title="pysegcnn.core.layers.Encoder" accesskey="n" rel="next">Next <span class="fa fa-arrow-circle-right"></span></a>
+
+
+ <a href="pysegcnn.core.layers.EncoderBlock.html" class="btn btn-neutral float-left" title="pysegcnn.core.layers.EncoderBlock" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left"></span> Previous</a>
+
+ </div>
+
<hr/>
diff --git a/docs/_build/html/source/generated/pysegcnn.core.layers.Encoder.html b/docs/_build/html/source/generated/pysegcnn.core.layers.Encoder.html
index fbc6521dd990ec09f97450ecf577d3c43672e8ce..ebc77d2b75706b774a48d7e646c2ba557eed174b 100644
--- a/docs/_build/html/source/generated/pysegcnn.core.layers.Encoder.html
+++ b/docs/_build/html/source/generated/pysegcnn.core.layers.Encoder.html
@@ -35,7 +35,9 @@
<link rel="index" title="Index" href="../../genindex.html" />
- <link rel="search" title="Search" href="../../search.html" />
+ <link rel="search" title="Search" href="../../search.html" />
+ <link rel="next" title="pysegcnn.core.layers.Decoder" href="pysegcnn.core.layers.Decoder.html" />
+ <link rel="prev" title="pysegcnn.core.layers.DecoderBlock" href="pysegcnn.core.layers.DecoderBlock.html" />
</head>
<body class="wy-body-for-nav">
@@ -80,10 +82,25 @@
- <ul>
+ <ul class="current">
<li class="toctree-l1"><a class="reference internal" href="../requirements.html">Requirements</a></li>
<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../api.html">API Reference</a></li>
+<li class="toctree-l1 current"><a class="reference internal" href="../api.html">API Reference</a><ul class="current">
+<li class="toctree-l2"><a class="reference internal" href="../api.html#dataset">Dataset</a></li>
+<li class="toctree-l2 current"><a class="reference internal" href="../api.html#models">Models</a><ul class="current">
+<li class="toctree-l3"><a class="reference internal" href="../api.html#layers">Layers</a></li>
+<li class="toctree-l3 current"><a class="reference internal" href="../api.html#encoder-decoder-architechture">Encoder-Decoder architechture</a><ul class="current">
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.EncoderBlock.html">pysegcnn.core.layers.EncoderBlock</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.DecoderBlock.html">pysegcnn.core.layers.DecoderBlock</a></li>
+<li class="toctree-l4 current"><a class="current reference internal" href="#">pysegcnn.core.layers.Encoder</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.Decoder.html">pysegcnn.core.layers.Decoder</a></li>
+</ul>
+</li>
+<li class="toctree-l3"><a class="reference internal" href="../api.html#neural-networks">Neural Networks</a></li>
+</ul>
+</li>
+</ul>
+</li>
</ul>
@@ -130,6 +147,8 @@
<li><a href="../../index.html" class="icon icon-home"></a> »</li>
+ <li><a href="../api.html">API Reference</a> »</li>
+
<li>pysegcnn.core.layers.Encoder</li>
@@ -156,148 +175,56 @@
<em class="property">class </em><code class="sig-prename descclassname">pysegcnn.core.layers.</code><code class="sig-name descname">Encoder</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">filters</span></em>, <em class="sig-param"><span class="n">block</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/layers.html#Encoder"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.layers.Encoder" title="Permalink to this definition">¶</a></dt>
<dd><p>Generic convolutional encoder.</p>
<p>When instanciating an encoder-decoder architechure, <code class="docutils literal notranslate"><span class="pre">filters</span></code> should be
-the same for <cite>pysegcnn.core.layers.Encoder</cite> and
-<cite>pysegcnn.core.layers.Decoder</cite>.</p>
-<p>See <cite>pysegcnn.core.models.UNet</cite> for an example implementation.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>filters</strong> (<cite>list</cite> [<cite>int</cite>]) – List of input channels to each convolutional block. The length of
-<code class="docutils literal notranslate"><span class="pre">filters</span></code> determines the depth of the encoder. The first element of
-<code class="docutils literal notranslate"><span class="pre">filters</span></code> has to be the number of channels of the input images.</p></li>
-<li><p><strong>block</strong> (<cite>pysegcnn.core.layers.EncoderBlock</cite>) – The convolutional block defining a layer in the encoder.
-A subclass of <cite>pysegcnn.core.layers.EncoderBlock</cite>, e.g.
-<cite>pysegcnn.core.layers.ConvBnReluMaxPool</cite>.</p></li>
-<li><p><strong>**kwargs</strong> (‘dict’ [<cite>str</cite>]) – Additional arguments passed to <cite>pysegcnn.core.layers.Conv2dSame</cite>.</p></li>
-</ul>
+the same for <a class="reference internal" href="#pysegcnn.core.layers.Encoder" title="pysegcnn.core.layers.Encoder"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.Encoder</span></code></a> and
+<a class="reference internal" href="pysegcnn.core.layers.Decoder.html#pysegcnn.core.layers.Decoder" title="pysegcnn.core.layers.Decoder"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.Decoder</span></code></a>.</p>
+<p>See <a class="reference internal" href="pysegcnn.core.models.UNet.html#pysegcnn.core.models.UNet" title="pysegcnn.core.models.UNet"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.models.UNet</span></code></a> for an example implementation.</p>
+<dl class="field-list">
+<dt class="field-odd">Attributes</dt>
+<dd class="field-odd"><dl>
+<dt><strong>features</strong><span class="classifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">numpy.ndarray</span></code></span></dt><dd><p>Input channels to each convolutional block, i.e. <code class="docutils literal notranslate"><span class="pre">filters</span></code>.</p>
+</dd>
+<dt><strong>block</strong><span class="classifier"><a class="reference internal" href="pysegcnn.core.layers.EncoderBlock.html#pysegcnn.core.layers.EncoderBlock" title="pysegcnn.core.layers.EncoderBlock"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.EncoderBlock</span></code></a></span></dt><dd><p>The convolutional block defining a layer in the encoder.
+A subclass of <a class="reference internal" href="pysegcnn.core.layers.EncoderBlock.html#pysegcnn.core.layers.EncoderBlock" title="pysegcnn.core.layers.EncoderBlock"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.EncoderBlock</span></code></a>, e.g.
+<a class="reference internal" href="pysegcnn.core.layers.ConvBnReluMaxPool.html#pysegcnn.core.layers.ConvBnReluMaxPool" title="pysegcnn.core.layers.ConvBnReluMaxPool"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.ConvBnReluMaxPool</span></code></a>.</p>
</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
+<dt><strong>layers</strong><span class="classifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">torch.nn.ModuleList</span></code></span></dt><dd><p>List of blocks in the encoder.</p>
</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
+<dt><strong>cache</strong><span class="classifier"><cite>dict</cite></span></dt><dd><dl class="simple">
+<dt>Intermediate encoder outputs. Dictionary with keys:</dt><dd><dl class="simple">
+<dt><code class="docutils literal notranslate"><span class="pre">'feature'</span></code></dt><dd><p>The intermediate encoder outputs (<code class="xref py py-class docutils literal notranslate"><span class="pre">torch.Tensor</span></code>).</p>
+</dd>
+<dt><code class="docutils literal notranslate"><span class="pre">'indices'</span></code></dt><dd><p>The indices of the max pooling layer, if required
+(<code class="xref py py-class docutils literal notranslate"><span class="pre">torch.Tensor</span></code>).</p>
+</dd>
+</dl>
+</dd>
+</dl>
+</dd>
+</dl>
</dd>
</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
-<td><p>Adds a child module to the current module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">apply</span></code>(fn)</p></td>
-<td><p>Applies <code class="docutils literal notranslate"><span class="pre">fn</span></code> recursively to every submodule (as returned by <code class="docutils literal notranslate"><span class="pre">.children()</span></code>) as well as self.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">bfloat16</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">bfloat16</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">buffers</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cpu</span></code>()</p></td>
-<td><p>Moves all model parameters and buffers to the CPU.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cuda</span></code>([device])</p></td>
-<td><p>Moves all model parameters and buffers to the GPU.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">double</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">double</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">eval</span></code>()</p></td>
-<td><p>Sets the module in evaluation mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">extra_repr</span></code>()</p></td>
-<td><p>Set the extra representation of the module</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">float</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to float datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward</span></code>(x)</p></td>
-<td><p>Forward pass of the encoder.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">half</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">half</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_state_dict</span></code>(state_dict[, strict])</p></td>
-<td><p>Copies parameters and buffers from <code class="xref py py-attr docutils literal notranslate"><span class="pre">state_dict</span></code> into this module and its descendants.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">modules</span></code>()</p></td>
-<td><p>Returns an iterator over all modules in the network.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_buffers</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_modules</span></code>([memo, prefix])</p></td>
-<td><p>Returns an iterator over all modules in the network, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_parameters</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parameters</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module parameters.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_backward_hook</span></code>(hook)</p></td>
-<td><p>Registers a backward hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_buffer</span></code>(name, tensor[, persistent])</p></td>
-<td><p>Adds a buffer to the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_pre_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward pre-hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_parameter</span></code>(name, param)</p></td>
-<td><p>Adds a parameter to the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">requires_grad_</span></code>([requires_grad])</p></td>
-<td><p>Change if autograd should record operations on parameters in this module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">state_dict</span></code>([destination, prefix, keep_vars])</p></td>
-<td><p>Returns a dictionary containing a whole state of the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to</span></code>(*args, **kwargs)</p></td>
-<td><p>Moves and/or casts the parameters and buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">train</span></code>([mode])</p></td>
-<td><p>Sets the module in training mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">type</span></code>(dst_type)</p></td>
-<td><p>Casts all parameters and buffers to <code class="xref py py-attr docutils literal notranslate"><span class="pre">dst_type</span></code>.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">zero_grad</span></code>()</p></td>
-<td><p>Sets gradients of all model parameters to zero.</p></td>
-</tr>
-</tbody>
-</table>
-<table class="docutils align-default">
-<colgroup>
-<col style="width: 62%" />
-<col style="width: 38%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><strong>__call__</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-even"><td><p><strong>share_memory</strong></p></td>
-<td></td>
-</tr>
-</tbody>
-</table>
<dl class="py method">
<dt id="pysegcnn.core.layers.Encoder.__init__">
<code class="sig-name descname">__init__</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">filters</span></em>, <em class="sig-param"><span class="n">block</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/layers.html#Encoder.__init__"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.layers.Encoder.__init__" title="Permalink to this definition">¶</a></dt>
-<dd><p>Initializes internal Module state, shared by both nn.Module and ScriptModule.</p>
+<dd><p>Initialize.</p>
+<dl class="field-list">
+<dt class="field-odd">Parameters</dt>
+<dd class="field-odd"><dl>
+<dt><strong>filters</strong><span class="classifier"><cite>list</cite> [<cite>int</cite>]</span></dt><dd><p>List of input channels to each convolutional block. The length of
+<code class="docutils literal notranslate"><span class="pre">filters</span></code> determines the depth of the encoder. The first element
+of <code class="docutils literal notranslate"><span class="pre">filters</span></code> has to be the number of channels of the input
+images.</p>
+</dd>
+<dt><strong>block</strong><span class="classifier"><a class="reference internal" href="pysegcnn.core.layers.EncoderBlock.html#pysegcnn.core.layers.EncoderBlock" title="pysegcnn.core.layers.EncoderBlock"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.EncoderBlock</span></code></a></span></dt><dd><p>The convolutional block defining a layer in the encoder.
+A subclass of <a class="reference internal" href="pysegcnn.core.layers.EncoderBlock.html#pysegcnn.core.layers.EncoderBlock" title="pysegcnn.core.layers.EncoderBlock"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.EncoderBlock</span></code></a>, e.g.
+<a class="reference internal" href="pysegcnn.core.layers.ConvBnReluMaxPool.html#pysegcnn.core.layers.ConvBnReluMaxPool" title="pysegcnn.core.layers.ConvBnReluMaxPool"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.ConvBnReluMaxPool</span></code></a>.</p>
+</dd>
+<dt><strong>**kwargs: `dict` [`str`]</strong></dt><dd><p>Additional arguments passed to
+<a class="reference internal" href="pysegcnn.core.layers.Conv2dSame.html#pysegcnn.core.layers.Conv2dSame" title="pysegcnn.core.layers.Conv2dSame"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.Conv2dSame</span></code></a>.</p>
+</dd>
+</dl>
+</dd>
+</dl>
</dd></dl>
<p class="rubric">Methods</p>
@@ -308,7 +235,7 @@ A subclass of <cite>pysegcnn.core.layers.EncoderBlock</cite>, e.g.
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#pysegcnn.core.layers.Encoder.__init__" title="pysegcnn.core.layers.Encoder.__init__"><code class="xref py py-obj docutils literal notranslate"><span class="pre">__init__</span></code></a>(filters, block, **kwargs)</p></td>
-<td><p>Initializes internal Module state, shared by both nn.Module and ScriptModule.</p></td>
+<td><p>Initialize.</p></td>
</tr>
<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
<td><p>Adds a child module to the current module.</p></td>
@@ -433,6 +360,15 @@ A subclass of <cite>pysegcnn.core.layers.EncoderBlock</cite>, e.g.
</div>
<footer>
+ <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
+
+ <a href="pysegcnn.core.layers.Decoder.html" class="btn btn-neutral float-right" title="pysegcnn.core.layers.Decoder" accesskey="n" rel="next">Next <span class="fa fa-arrow-circle-right"></span></a>
+
+
+ <a href="pysegcnn.core.layers.DecoderBlock.html" class="btn btn-neutral float-left" title="pysegcnn.core.layers.DecoderBlock" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left"></span> Previous</a>
+
+ </div>
+
<hr/>
diff --git a/docs/_build/html/source/generated/pysegcnn.core.layers.EncoderBlock.html b/docs/_build/html/source/generated/pysegcnn.core.layers.EncoderBlock.html
index c41607a518f10ae7ad99df71739fae5a3efd3c27..8fb1f3d14550274f6d6d625211b6ced37438a236 100644
--- a/docs/_build/html/source/generated/pysegcnn.core.layers.EncoderBlock.html
+++ b/docs/_build/html/source/generated/pysegcnn.core.layers.EncoderBlock.html
@@ -35,7 +35,9 @@
<link rel="index" title="Index" href="../../genindex.html" />
- <link rel="search" title="Search" href="../../search.html" />
+ <link rel="search" title="Search" href="../../search.html" />
+ <link rel="next" title="pysegcnn.core.layers.DecoderBlock" href="pysegcnn.core.layers.DecoderBlock.html" />
+ <link rel="prev" title="pysegcnn.core.layers.ConvBnReluUpsample" href="pysegcnn.core.layers.ConvBnReluUpsample.html" />
</head>
<body class="wy-body-for-nav">
@@ -80,10 +82,25 @@
- <ul>
+ <ul class="current">
<li class="toctree-l1"><a class="reference internal" href="../requirements.html">Requirements</a></li>
<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../api.html">API Reference</a></li>
+<li class="toctree-l1 current"><a class="reference internal" href="../api.html">API Reference</a><ul class="current">
+<li class="toctree-l2"><a class="reference internal" href="../api.html#dataset">Dataset</a></li>
+<li class="toctree-l2 current"><a class="reference internal" href="../api.html#models">Models</a><ul class="current">
+<li class="toctree-l3"><a class="reference internal" href="../api.html#layers">Layers</a></li>
+<li class="toctree-l3 current"><a class="reference internal" href="../api.html#encoder-decoder-architechture">Encoder-Decoder architechture</a><ul class="current">
+<li class="toctree-l4 current"><a class="current reference internal" href="#">pysegcnn.core.layers.EncoderBlock</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.DecoderBlock.html">pysegcnn.core.layers.DecoderBlock</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.Encoder.html">pysegcnn.core.layers.Encoder</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.layers.Decoder.html">pysegcnn.core.layers.Decoder</a></li>
+</ul>
+</li>
+<li class="toctree-l3"><a class="reference internal" href="../api.html#neural-networks">Neural Networks</a></li>
+</ul>
+</li>
+</ul>
+</li>
</ul>
@@ -130,6 +147,8 @@
<li><a href="../../index.html" class="icon icon-home"></a> »</li>
+ <li><a href="../api.html">API Reference</a> »</li>
+
<li>pysegcnn.core.layers.EncoderBlock</li>
@@ -155,147 +174,30 @@
<dt id="pysegcnn.core.layers.EncoderBlock">
<em class="property">class </em><code class="sig-prename descclassname">pysegcnn.core.layers.</code><code class="sig-name descname">EncoderBlock</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">out_channels</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/layers.html#EncoderBlock"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.layers.EncoderBlock" title="Permalink to this definition">¶</a></dt>
<dd><p>Block of a convolutional encoder.</p>
-<dl class="field-list simple">
+<dl class="py method">
+<dt id="pysegcnn.core.layers.EncoderBlock.__init__">
+<code class="sig-name descname">__init__</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">out_channels</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/layers.html#EncoderBlock.__init__"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.layers.EncoderBlock.__init__" title="Permalink to this definition">¶</a></dt>
+<dd><p>Initialize.</p>
+<dl class="field-list">
<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>in_channels</strong> (<cite>int</cite>) – Number of input channels.</p></li>
-<li><p><strong>out_channels</strong> (<cite>int</cite>) – Number of output channels.</p></li>
-<li><p><strong>**kwargs</strong> (‘dict’ [<cite>str</cite>]) – Additional arguments passed to <cite>pysegcnn.core.layers.Conv2dSame</cite>.</p></li>
-</ul>
+<dd class="field-odd"><dl>
+<dt><strong>in_channels</strong><span class="classifier"><cite>int</cite></span></dt><dd><p>Number of input channels.</p>
</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
+<dt><strong>out_channels</strong><span class="classifier"><cite>int</cite></span></dt><dd><p>Number of output channels.</p>
</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
+<dt><strong>**kwargs: `dict` [`str`]</strong></dt><dd><p>Additional arguments passed to
+<a class="reference internal" href="pysegcnn.core.layers.Conv2dSame.html#pysegcnn.core.layers.Conv2dSame" title="pysegcnn.core.layers.Conv2dSame"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.Conv2dSame</span></code></a>.</p>
+</dd>
+</dl>
+</dd>
+<dt class="field-even">Raises</dt>
+<dd class="field-even"><dl class="simple">
+<dt>TypeError</dt><dd><p>Raised if <code class="xref py py-meth docutils literal notranslate"><span class="pre">layers()</span></code> method does
+not return an instance of <code class="xref py py-class docutils literal notranslate"><span class="pre">torch.nn.Sequential</span></code>.</p>
+</dd>
+</dl>
</dd>
</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
-<td><p>Adds a child module to the current module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">apply</span></code>(fn)</p></td>
-<td><p>Applies <code class="docutils literal notranslate"><span class="pre">fn</span></code> recursively to every submodule (as returned by <code class="docutils literal notranslate"><span class="pre">.children()</span></code>) as well as self.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">bfloat16</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">bfloat16</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">buffers</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cpu</span></code>()</p></td>
-<td><p>Moves all model parameters and buffers to the CPU.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cuda</span></code>([device])</p></td>
-<td><p>Moves all model parameters and buffers to the GPU.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">double</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">double</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">downsample</span></code>(x)</p></td>
-<td><p>Define the downsampling method.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">eval</span></code>()</p></td>
-<td><p>Sets the module in evaluation mode.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">extra_repr</span></code>()</p></td>
-<td><p>Set the extra representation of the module</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">float</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to float datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward</span></code>(x)</p></td>
-<td><p>Forward pass of an encoder block.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">half</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">half</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">layers</span></code>()</p></td>
-<td><p>Define the layers of the block.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_state_dict</span></code>(state_dict[, strict])</p></td>
-<td><p>Copies parameters and buffers from <code class="xref py py-attr docutils literal notranslate"><span class="pre">state_dict</span></code> into this module and its descendants.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">modules</span></code>()</p></td>
-<td><p>Returns an iterator over all modules in the network.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_buffers</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_modules</span></code>([memo, prefix])</p></td>
-<td><p>Returns an iterator over all modules in the network, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_parameters</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parameters</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module parameters.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_backward_hook</span></code>(hook)</p></td>
-<td><p>Registers a backward hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_buffer</span></code>(name, tensor[, persistent])</p></td>
-<td><p>Adds a buffer to the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_pre_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward pre-hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_parameter</span></code>(name, param)</p></td>
-<td><p>Adds a parameter to the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">requires_grad_</span></code>([requires_grad])</p></td>
-<td><p>Change if autograd should record operations on parameters in this module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">state_dict</span></code>([destination, prefix, keep_vars])</p></td>
-<td><p>Returns a dictionary containing a whole state of the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to</span></code>(*args, **kwargs)</p></td>
-<td><p>Moves and/or casts the parameters and buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">train</span></code>([mode])</p></td>
-<td><p>Sets the module in training mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">type</span></code>(dst_type)</p></td>
-<td><p>Casts all parameters and buffers to <code class="xref py py-attr docutils literal notranslate"><span class="pre">dst_type</span></code>.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">zero_grad</span></code>()</p></td>
-<td><p>Sets gradients of all model parameters to zero.</p></td>
-</tr>
-</tbody>
-</table>
-<table class="docutils align-default">
-<colgroup>
-<col style="width: 62%" />
-<col style="width: 38%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><strong>__call__</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-even"><td><p><strong>share_memory</strong></p></td>
-<td></td>
-</tr>
-</tbody>
-</table>
-<dl class="py method">
-<dt id="pysegcnn.core.layers.EncoderBlock.__init__">
-<code class="sig-name descname">__init__</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">out_channels</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/layers.html#EncoderBlock.__init__"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.layers.EncoderBlock.__init__" title="Permalink to this definition">¶</a></dt>
-<dd><p>Initializes internal Module state, shared by both nn.Module and ScriptModule.</p>
</dd></dl>
<p class="rubric">Methods</p>
@@ -306,7 +208,7 @@
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#pysegcnn.core.layers.EncoderBlock.__init__" title="pysegcnn.core.layers.EncoderBlock.__init__"><code class="xref py py-obj docutils literal notranslate"><span class="pre">__init__</span></code></a>(in_channels, out_channels, **kwargs)</p></td>
-<td><p>Initializes internal Module state, shared by both nn.Module and ScriptModule.</p></td>
+<td><p>Initialize.</p></td>
</tr>
<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
<td><p>Adds a child module to the current module.</p></td>
@@ -437,6 +339,15 @@
</div>
<footer>
+ <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
+
+ <a href="pysegcnn.core.layers.DecoderBlock.html" class="btn btn-neutral float-right" title="pysegcnn.core.layers.DecoderBlock" accesskey="n" rel="next">Next <span class="fa fa-arrow-circle-right"></span></a>
+
+
+ <a href="pysegcnn.core.layers.ConvBnReluUpsample.html" class="btn btn-neutral float-left" title="pysegcnn.core.layers.ConvBnReluUpsample" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left"></span> Previous</a>
+
+ </div>
+
<hr/>
diff --git a/docs/_build/html/source/generated/pysegcnn.core.layers.html b/docs/_build/html/source/generated/pysegcnn.core.layers.html
deleted file mode 100644
index 5f954e54df2408834d52277904eab6e20c00ff52..0000000000000000000000000000000000000000
--- a/docs/_build/html/source/generated/pysegcnn.core.layers.html
+++ /dev/null
@@ -1,283 +0,0 @@
-
-
-<!DOCTYPE html>
-<html class="writer-html5" lang="en" >
-<head>
- <meta charset="utf-8">
-
- <meta name="viewport" content="width=device-width, initial-scale=1.0">
-
- <title>pysegcnn.core.layers — PySegCNN documentation</title>
-
-
-
- <link rel="stylesheet" href="../../_static/css/theme.css" type="text/css" />
- <link rel="stylesheet" href="../../_static/pygments.css" type="text/css" />
-
-
-
-
-
-
-
- <!--[if lt IE 9]>
- <script src="../../_static/js/html5shiv.min.js"></script>
- <![endif]-->
-
-
- <script type="text/javascript" id="documentation_options" data-url_root="../../" src="../../_static/documentation_options.js"></script>
- <script src="../../_static/jquery.js"></script>
- <script src="../../_static/underscore.js"></script>
- <script src="../../_static/doctools.js"></script>
- <script src="../../_static/language_data.js"></script>
-
- <script type="text/javascript" src="../../_static/js/theme.js"></script>
-
-
- <link rel="index" title="Index" href="../../genindex.html" />
- <link rel="search" title="Search" href="../../search.html" />
- <link rel="next" title="pysegcnn.core.models" href="pysegcnn.core.models.html" />
- <link rel="prev" title="API Reference" href="../api.html" />
-</head>
-
-<body class="wy-body-for-nav">
-
-
- <div class="wy-grid-for-nav">
-
- <nav data-toggle="wy-nav-shift" class="wy-nav-side">
- <div class="wy-side-scroll">
- <div class="wy-side-nav-search" >
-
-
-
- <a href="../../index.html" class="icon icon-home" alt="Documentation Home"> PySegCNN
-
-
-
- </a>
-
-
-
-
-
-
-
-<div role="search">
- <form id="rtd-search-form" class="wy-form" action="../../search.html" method="get">
- <input type="text" name="q" placeholder="Search docs" />
- <input type="hidden" name="check_keywords" value="yes" />
- <input type="hidden" name="area" value="default" />
- </form>
-</div>
-
-
- </div>
-
-
- <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
-
-
-
-
-
-
- <ul class="current">
-<li class="toctree-l1"><a class="reference internal" href="../requirements.html">Requirements</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
-<li class="toctree-l1 current"><a class="reference internal" href="../api.html">API Reference</a><ul class="current">
-<li class="toctree-l2 current"><a class="reference internal" href="../api.html#layers-and-models">Layers and Models</a><ul class="current">
-<li class="toctree-l3 current"><a class="current reference internal" href="#">pysegcnn.core.layers</a><ul>
-<li class="toctree-l4"><a class="reference internal" href="#license">License</a></li>
-</ul>
-</li>
-<li class="toctree-l3"><a class="reference internal" href="pysegcnn.core.models.html">pysegcnn.core.models</a></li>
-</ul>
-</li>
-</ul>
-</li>
-</ul>
-
-
-
- </div>
-
- </div>
- </nav>
-
- <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
-
-
- <nav class="wy-nav-top" aria-label="top navigation">
-
- <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
- <a href="../../index.html">PySegCNN</a>
-
- </nav>
-
-
- <div class="wy-nav-content">
-
- <div class="rst-content">
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-<div role="navigation" aria-label="breadcrumbs navigation">
-
- <ul class="wy-breadcrumbs">
-
- <li><a href="../../index.html" class="icon icon-home"></a> »</li>
-
- <li><a href="../api.html">API Reference</a> »</li>
-
- <li>pysegcnn.core.layers</li>
-
-
- <li class="wy-breadcrumbs-aside">
-
-
- <a href="../../_sources/source/generated/pysegcnn.core.layers.rst.txt" rel="nofollow"> View page source</a>
-
-
- </li>
-
- </ul>
-
-
- <hr/>
-</div>
- <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
- <div itemprop="articleBody">
-
- <div class="section" id="module-pysegcnn.core.layers">
-<span id="pysegcnn-core-layers"></span><h1>pysegcnn.core.layers<a class="headerlink" href="#module-pysegcnn.core.layers" title="Permalink to this headline">¶</a></h1>
-<p>Layers of a convolutional encoder-decoder network.</p>
-<div class="section" id="license">
-<h2>License<a class="headerlink" href="#license" title="Permalink to this headline">¶</a></h2>
-<blockquote>
-<div><p>Copyright (c) 2020 Daniel Frisinghelli</p>
-<p>This source code is licensed under the GNU General Public License v3.</p>
-<p>See the LICENSE file in the repository’s root directory.</p>
-</div></blockquote>
-<p class="rubric">Functions</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">conv_bn_relu</span></code>(in_channels, out_channels, **kwargs)</p></td>
-<td><p>Block of convolution, batch normalization and rectified linear unit.</p></td>
-</tr>
-</tbody>
-</table>
-<p class="rubric">Classes</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">Block</span></code>(in_channels, out_channels, **kwargs)</p></td>
-<td><p>Basic convolutional block.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">Conv2dSame</span></code>(*args, **kwargs)</p></td>
-<td><p>A convolution preserving the shape of its input.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">ConvBnReluMaxPool</span></code>(in_channels, out_channels, …)</p></td>
-<td><p>Block of convolution, batchnorm, relu and 2x2 max pool.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">ConvBnReluMaxUnpool</span></code>(in_channels, …)</p></td>
-<td><p>Block of convolution, batchnorm, relu and 2x2 max unpool.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">ConvBnReluUpsample</span></code>(in_channels, …)</p></td>
-<td><p>Block of convolution, batchnorm, relu and nearest neighbor upsampling.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">Decoder</span></code>(filters, block[, skip])</p></td>
-<td><p>Generic convolutional decoder.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">DecoderBlock</span></code>(in_channels, out_channels, **kwargs)</p></td>
-<td><p>Block of a convolutional decoder.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">Encoder</span></code>(filters, block, **kwargs)</p></td>
-<td><p>Generic convolutional encoder.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">EncoderBlock</span></code>(in_channels, out_channels, **kwargs)</p></td>
-<td><p>Block of a convolutional encoder.</p></td>
-</tr>
-</tbody>
-</table>
-</div>
-</div>
-
-
- </div>
-
- </div>
- <footer>
-
- <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
-
- <a href="pysegcnn.core.models.html" class="btn btn-neutral float-right" title="pysegcnn.core.models" accesskey="n" rel="next">Next <span class="fa fa-arrow-circle-right"></span></a>
-
-
- <a href="../api.html" class="btn btn-neutral float-left" title="API Reference" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left"></span> Previous</a>
-
- </div>
-
-
- <hr/>
-
- <div role="contentinfo">
- <p>
-
- © Copyright 2020, Daniel Frisinghelli
-
- </p>
- </div>
-
-
-
- Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a
-
- <a href="https://github.com/rtfd/sphinx_rtd_theme">theme</a>
-
- provided by <a href="https://readthedocs.org">Read the Docs</a>.
-
-</footer>
-
- </div>
- </div>
-
- </section>
-
- </div>
-
-
- <script type="text/javascript">
- jQuery(function () {
- SphinxRtdTheme.Navigation.enable(true);
- });
- </script>
-
-
-
-
-
-
-</body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/source/generated/pysegcnn.core.models.Network.html b/docs/_build/html/source/generated/pysegcnn.core.models.Network.html
index 1147eae92a395b1a5d42f77e52f5004efe4ac674..d5fb8006a7f6568df408341c98e6817d356a4dbf 100644
--- a/docs/_build/html/source/generated/pysegcnn.core.models.Network.html
+++ b/docs/_build/html/source/generated/pysegcnn.core.models.Network.html
@@ -35,7 +35,9 @@
<link rel="index" title="Index" href="../../genindex.html" />
- <link rel="search" title="Search" href="../../search.html" />
+ <link rel="search" title="Search" href="../../search.html" />
+ <link rel="next" title="pysegcnn.core.models.UNet" href="pysegcnn.core.models.UNet.html" />
+ <link rel="prev" title="pysegcnn.core.layers.Decoder" href="pysegcnn.core.layers.Decoder.html" />
</head>
<body class="wy-body-for-nav">
@@ -80,10 +82,23 @@
- <ul>
+ <ul class="current">
<li class="toctree-l1"><a class="reference internal" href="../requirements.html">Requirements</a></li>
<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../api.html">API Reference</a></li>
+<li class="toctree-l1 current"><a class="reference internal" href="../api.html">API Reference</a><ul class="current">
+<li class="toctree-l2"><a class="reference internal" href="../api.html#dataset">Dataset</a></li>
+<li class="toctree-l2 current"><a class="reference internal" href="../api.html#models">Models</a><ul class="current">
+<li class="toctree-l3"><a class="reference internal" href="../api.html#layers">Layers</a></li>
+<li class="toctree-l3"><a class="reference internal" href="../api.html#encoder-decoder-architechture">Encoder-Decoder architechture</a></li>
+<li class="toctree-l3 current"><a class="reference internal" href="../api.html#neural-networks">Neural Networks</a><ul class="current">
+<li class="toctree-l4 current"><a class="current reference internal" href="#">pysegcnn.core.models.Network</a></li>
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.models.UNet.html">pysegcnn.core.models.UNet</a></li>
+</ul>
+</li>
+</ul>
+</li>
+</ul>
+</li>
</ul>
@@ -130,6 +145,8 @@
<li><a href="../../index.html" class="icon icon-home"></a> »</li>
+ <li><a href="../api.html">API Reference</a> »</li>
+
<li>pysegcnn.core.models.Network</li>
@@ -156,153 +173,19 @@
<em class="property">class </em><code class="sig-prename descclassname">pysegcnn.core.models.</code><code class="sig-name descname">Network</code><a class="reference internal" href="../../_modules/pysegcnn/core/models.html#Network"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.models.Network" title="Permalink to this definition">¶</a></dt>
<dd><p>Generic Network class.</p>
<p>The base class for each model. If you want to implement a new model,
-inherit the <code class="docutils literal notranslate"><span class="pre">~pysegcnn.core.models.Network</span></code> class.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
-<dd class="field-odd"><p></p>
-</dd>
-<dt class="field-even">Return type</dt>
-<dd class="field-even"><p>None.</p>
-</dd>
+inherit the <a class="reference internal" href="#pysegcnn.core.models.Network" title="pysegcnn.core.models.Network"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.models.Network</span></code></a> class.</p>
+<dl class="field-list">
<dt class="field-odd">Attributes</dt>
-<dd class="field-odd"><dl class="simple">
-<dt><code class="xref py py-obj docutils literal notranslate"><span class="pre">state</span></code></dt><dd><p>Return the model state file.</p>
+<dd class="field-odd"><dl>
+<dt><strong>state_file</strong><span class="classifier"><cite>str</cite> or <cite>None</cite> or <code class="xref py py-class docutils literal notranslate"><span class="pre">pathlib.Path</span></code></span></dt><dd><p>The model state file, where the model parameters are saved.</p>
</dd>
</dl>
</dd>
</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
-<td><p>Adds a child module to the current module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">apply</span></code>(fn)</p></td>
-<td><p>Applies <code class="docutils literal notranslate"><span class="pre">fn</span></code> recursively to every submodule (as returned by <code class="docutils literal notranslate"><span class="pre">.children()</span></code>) as well as self.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">bfloat16</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">bfloat16</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">buffers</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cpu</span></code>()</p></td>
-<td><p>Moves all model parameters and buffers to the CPU.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cuda</span></code>([device])</p></td>
-<td><p>Moves all model parameters and buffers to the GPU.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">double</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">double</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">eval</span></code>()</p></td>
-<td><p>Sets the module in evaluation mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">extra_repr</span></code>()</p></td>
-<td><p>Set the extra representation of the module</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">float</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to float datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward</span></code>(*input)</p></td>
-<td><p></p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">freeze</span></code>()</p></td>
-<td><p>Freeze the weights of a model.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">half</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">half</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">load</span></code>(state_file)</p></td>
-<td><p>Load a model state.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_state_dict</span></code>(state_dict[, strict])</p></td>
-<td><p>Copies parameters and buffers from <code class="xref py py-attr docutils literal notranslate"><span class="pre">state_dict</span></code> into this module and its descendants.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">modules</span></code>()</p></td>
-<td><p>Returns an iterator over all modules in the network.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_buffers</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_modules</span></code>([memo, prefix])</p></td>
-<td><p>Returns an iterator over all modules in the network, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_parameters</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parameters</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module parameters.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_backward_hook</span></code>(hook)</p></td>
-<td><p>Registers a backward hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_buffer</span></code>(name, tensor[, persistent])</p></td>
-<td><p>Adds a buffer to the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_pre_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward pre-hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_parameter</span></code>(name, param)</p></td>
-<td><p>Adds a parameter to the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">requires_grad_</span></code>([requires_grad])</p></td>
-<td><p>Change if autograd should record operations on parameters in this module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">save</span></code>(state_file, optimizer[, bands])</p></td>
-<td><p>Save the model state.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">state_dict</span></code>([destination, prefix, keep_vars])</p></td>
-<td><p>Returns a dictionary containing a whole state of the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to</span></code>(*args, **kwargs)</p></td>
-<td><p>Moves and/or casts the parameters and buffers.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">train</span></code>([mode])</p></td>
-<td><p>Sets the module in training mode.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">type</span></code>(dst_type)</p></td>
-<td><p>Casts all parameters and buffers to <code class="xref py py-attr docutils literal notranslate"><span class="pre">dst_type</span></code>.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">unfreeze</span></code>()</p></td>
-<td><p>Unfreeze the weights of a model.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">zero_grad</span></code>()</p></td>
-<td><p>Sets gradients of all model parameters to zero.</p></td>
-</tr>
-</tbody>
-</table>
-<table class="docutils align-default">
-<colgroup>
-<col style="width: 62%" />
-<col style="width: 38%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><strong>__call__</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-even"><td><p><strong>share_memory</strong></p></td>
-<td></td>
-</tr>
-</tbody>
-</table>
<dl class="py method">
<dt id="pysegcnn.core.models.Network.__init__">
<code class="sig-name descname">__init__</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/models.html#Network.__init__"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.models.Network.__init__" title="Permalink to this definition">¶</a></dt>
-<dd><p>Initializes internal Module state, shared by both nn.Module and ScriptModule.</p>
+<dd><p>Initialize.</p>
</dd></dl>
<p class="rubric">Methods</p>
@@ -313,7 +196,7 @@ inherit the <code class="docutils literal notranslate"><span class="pre">~pysegc
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#pysegcnn.core.models.Network.__init__" title="pysegcnn.core.models.Network.__init__"><code class="xref py py-obj docutils literal notranslate"><span class="pre">__init__</span></code></a>()</p></td>
-<td><p>Initializes internal Module state, shared by both nn.Module and ScriptModule.</p></td>
+<td><p>Initialize.</p></td>
</tr>
<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
<td><p>Adds a child module to the current module.</p></td>
@@ -453,6 +336,15 @@ inherit the <code class="docutils literal notranslate"><span class="pre">~pysegc
</div>
<footer>
+ <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
+
+ <a href="pysegcnn.core.models.UNet.html" class="btn btn-neutral float-right" title="pysegcnn.core.models.UNet" accesskey="n" rel="next">Next <span class="fa fa-arrow-circle-right"></span></a>
+
+
+ <a href="pysegcnn.core.layers.Decoder.html" class="btn btn-neutral float-left" title="pysegcnn.core.layers.Decoder" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left"></span> Previous</a>
+
+ </div>
+
<hr/>
diff --git a/docs/_build/html/source/generated/pysegcnn.core.models.UNet.html b/docs/_build/html/source/generated/pysegcnn.core.models.UNet.html
index b276d7b9a83f8b8fa17bf0168951c86df3fb5a31..2a81325512b299c21c9b09a50b54c815b6461aa9 100644
--- a/docs/_build/html/source/generated/pysegcnn.core.models.UNet.html
+++ b/docs/_build/html/source/generated/pysegcnn.core.models.UNet.html
@@ -35,7 +35,8 @@
<link rel="index" title="Index" href="../../genindex.html" />
- <link rel="search" title="Search" href="../../search.html" />
+ <link rel="search" title="Search" href="../../search.html" />
+ <link rel="prev" title="pysegcnn.core.models.Network" href="pysegcnn.core.models.Network.html" />
</head>
<body class="wy-body-for-nav">
@@ -80,10 +81,23 @@
- <ul>
+ <ul class="current">
<li class="toctree-l1"><a class="reference internal" href="../requirements.html">Requirements</a></li>
<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../api.html">API Reference</a></li>
+<li class="toctree-l1 current"><a class="reference internal" href="../api.html">API Reference</a><ul class="current">
+<li class="toctree-l2"><a class="reference internal" href="../api.html#dataset">Dataset</a></li>
+<li class="toctree-l2 current"><a class="reference internal" href="../api.html#models">Models</a><ul class="current">
+<li class="toctree-l3"><a class="reference internal" href="../api.html#layers">Layers</a></li>
+<li class="toctree-l3"><a class="reference internal" href="../api.html#encoder-decoder-architechture">Encoder-Decoder architechture</a></li>
+<li class="toctree-l3 current"><a class="reference internal" href="../api.html#neural-networks">Neural Networks</a><ul class="current">
+<li class="toctree-l4"><a class="reference internal" href="pysegcnn.core.models.Network.html">pysegcnn.core.models.Network</a></li>
+<li class="toctree-l4 current"><a class="current reference internal" href="#">pysegcnn.core.models.UNet</a></li>
+</ul>
+</li>
+</ul>
+</li>
+</ul>
+</li>
</ul>
@@ -130,6 +144,8 @@
<li><a href="../../index.html" class="icon icon-home"></a> »</li>
+ <li><a href="../api.html">API Reference</a> »</li>
+
<li>pysegcnn.core.models.UNet</li>
@@ -154,170 +170,60 @@
<dl class="py class">
<dt id="pysegcnn.core.models.UNet">
<em class="property">class </em><code class="sig-prename descclassname">pysegcnn.core.models.</code><code class="sig-name descname">UNet</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">nclasses</span></em>, <em class="sig-param"><span class="n">filters</span></em>, <em class="sig-param"><span class="n">skip</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/models.html#UNet"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.models.UNet" title="Permalink to this definition">¶</a></dt>
-<dd><p>A PyTorch implementation of <a class="reference external" href="https://arxiv.org/abs/1505.04597">U-Net</a>.</p>
-<dl class="simple">
-<dt>Slightly modified version of U-Net:</dt><dd><ul class="simple">
-<li><p>each convolution is followed by a batch normalization layer</p></li>
-<li><p>the upsampling is implemented by a 2x2 max unpooling operation</p></li>
+<dd><p>A slightly modified implementation of <a class="reference external" href="https://arxiv.org/abs/1505.04597">U-Net</a> in PyTorch.</p>
+<div class="admonition important">
+<p class="admonition-title">Important</p>
+<ul class="simple">
+<li><p>Each convolution is followed by a batch normalization layer</p></li>
+<li><p>Upsampling is implemented by a 2x2 max unpooling operation</p></li>
</ul>
+</div>
+<dl class="field-list">
+<dt class="field-odd">Attributes</dt>
+<dd class="field-odd"><dl>
+<dt><strong>in_channels</strong><span class="classifier"><cite>int</cite></span></dt><dd><p>Number of channels of the input images.</p>
</dd>
-</dl>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>in_channels</strong> (<cite>int</cite>) – Number of channels of the input images.</p></li>
-<li><p><strong>nclasses</strong> (<cite>int</cite>) – Number of classes.</p></li>
-<li><p><strong>filters</strong> (<cite>list</cite> [<cite>int</cite>]) – List of input channels to each convolutional block.</p></li>
-<li><p><strong>skip</strong> (<cite>bool</cite>) – Whether to apply skip connections from the encoder to the decoder.</p></li>
-<li><p><strong>**kwargs</strong> (‘dict’ [<cite>str</cite>]) – Additional keyword arguments passed to
-<cite>pysegcnn.core.layers.Conv2dSame</cite>.</p></li>
-</ul>
+<dt><strong>nclasses</strong><span class="classifier"><cite>int</cite></span></dt><dd><p>Number of classes.</p>
</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p></p>
+<dt><strong>kwargs</strong><span class="classifier"><cite>dict</cite> [<cite>str</cite>]</span></dt><dd><p>Additional keyword arguments passed to
+<a class="reference internal" href="pysegcnn.core.layers.Conv2dSame.html#pysegcnn.core.layers.Conv2dSame" title="pysegcnn.core.layers.Conv2dSame"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.Conv2dSame</span></code></a>.</p>
</dd>
-<dt class="field-odd">Return type</dt>
-<dd class="field-odd"><p>None.</p>
+<dt><strong>nfilters</strong><span class="classifier"><cite>list</cite> [<cite>int</cite>]</span></dt><dd><p>List of input channels to each convolutional block.</p>
</dd>
-<dt class="field-even">Attributes</dt>
-<dd class="field-even"><dl class="simple">
-<dt><code class="xref py py-obj docutils literal notranslate"><span class="pre">state</span></code></dt><dd><p>Return the model state file.</p>
+<dt><strong>skip</strong><span class="classifier"><cite>bool</cite></span></dt><dd><p>Whether to apply skip connections from the encoder to the decoder.</p>
+</dd>
+<dt><strong>epoch</strong><span class="classifier"><cite>int</cite></span></dt><dd><p>Number of epochs the model was trained.</p>
+</dd>
+<dt><strong>encoder</strong><span class="classifier"><a class="reference internal" href="pysegcnn.core.layers.Encoder.html#pysegcnn.core.layers.Encoder" title="pysegcnn.core.layers.Encoder"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.Encoder</span></code></a></span></dt><dd><p>The convolutional encoder.</p>
+</dd>
+<dt><strong>decoder</strong><span class="classifier"><a class="reference internal" href="pysegcnn.core.layers.Decoder.html#pysegcnn.core.layers.Decoder" title="pysegcnn.core.layers.Decoder"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.Decoder</span></code></a></span></dt><dd><p>The convolutional decoder.</p>
+</dd>
+<dt><strong>classifier</strong><span class="classifier"><a class="reference internal" href="pysegcnn.core.layers.Conv2dSame.html#pysegcnn.core.layers.Conv2dSame" title="pysegcnn.core.layers.Conv2dSame"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.Conv2dSame</span></code></a></span></dt><dd><p>The classification layer, a 1x1 convolution.</p>
</dd>
</dl>
</dd>
</dl>
-<p class="rubric">Methods</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
-<td><p>Adds a child module to the current module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">apply</span></code>(fn)</p></td>
-<td><p>Applies <code class="docutils literal notranslate"><span class="pre">fn</span></code> recursively to every submodule (as returned by <code class="docutils literal notranslate"><span class="pre">.children()</span></code>) as well as self.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">bfloat16</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">bfloat16</span></code> datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">buffers</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module buffers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cpu</span></code>()</p></td>
-<td><p>Moves all model parameters and buffers to the CPU.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">cuda</span></code>([device])</p></td>
-<td><p>Moves all model parameters and buffers to the GPU.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">double</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">double</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">eval</span></code>()</p></td>
-<td><p>Sets the module in evaluation mode.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">extra_repr</span></code>()</p></td>
-<td><p>Set the extra representation of the module</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">float</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to float datatype.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward</span></code>(x)</p></td>
-<td><p>Forward propagation of U-Net.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">freeze</span></code>()</p></td>
-<td><p>Freeze the weights of a model.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">half</span></code>()</p></td>
-<td><p>Casts all floating point parameters and buffers to <code class="docutils literal notranslate"><span class="pre">half</span></code> datatype.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">load</span></code>(state_file)</p></td>
-<td><p>Load a model state.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_state_dict</span></code>(state_dict[, strict])</p></td>
-<td><p>Copies parameters and buffers from <code class="xref py py-attr docutils literal notranslate"><span class="pre">state_dict</span></code> into this module and its descendants.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">modules</span></code>()</p></td>
-<td><p>Returns an iterator over all modules in the network.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_buffers</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_children</span></code>()</p></td>
-<td><p>Returns an iterator over immediate children modules, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_modules</span></code>([memo, prefix])</p></td>
-<td><p>Returns an iterator over all modules in the network, yielding both the name of the module as well as the module itself.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">named_parameters</span></code>([prefix, recurse])</p></td>
-<td><p>Returns an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">parameters</span></code>([recurse])</p></td>
-<td><p>Returns an iterator over module parameters.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_backward_hook</span></code>(hook)</p></td>
-<td><p>Registers a backward hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_buffer</span></code>(name, tensor[, persistent])</p></td>
-<td><p>Adds a buffer to the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward hook on the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_forward_pre_hook</span></code>(hook)</p></td>
-<td><p>Registers a forward pre-hook on the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">register_parameter</span></code>(name, param)</p></td>
-<td><p>Adds a parameter to the module.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">requires_grad_</span></code>([requires_grad])</p></td>
-<td><p>Change if autograd should record operations on parameters in this module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">save</span></code>(state_file, optimizer[, bands])</p></td>
-<td><p>Save the model state.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">state_dict</span></code>([destination, prefix, keep_vars])</p></td>
-<td><p>Returns a dictionary containing a whole state of the module.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">to</span></code>(*args, **kwargs)</p></td>
-<td><p>Moves and/or casts the parameters and buffers.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">train</span></code>([mode])</p></td>
-<td><p>Sets the module in training mode.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">type</span></code>(dst_type)</p></td>
-<td><p>Casts all parameters and buffers to <code class="xref py py-attr docutils literal notranslate"><span class="pre">dst_type</span></code>.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">unfreeze</span></code>()</p></td>
-<td><p>Unfreeze the weights of a model.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">zero_grad</span></code>()</p></td>
-<td><p>Sets gradients of all model parameters to zero.</p></td>
-</tr>
-</tbody>
-</table>
-<table class="docutils align-default">
-<colgroup>
-<col style="width: 62%" />
-<col style="width: 38%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><strong>__call__</strong></p></td>
-<td></td>
-</tr>
-<tr class="row-even"><td><p><strong>share_memory</strong></p></td>
-<td></td>
-</tr>
-</tbody>
-</table>
<dl class="py method">
<dt id="pysegcnn.core.models.UNet.__init__">
<code class="sig-name descname">__init__</code><span class="sig-paren">(</span><em class="sig-param"><span class="n">in_channels</span></em>, <em class="sig-param"><span class="n">nclasses</span></em>, <em class="sig-param"><span class="n">filters</span></em>, <em class="sig-param"><span class="n">skip</span></em>, <em class="sig-param"><span class="o">**</span><span class="n">kwargs</span></em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/pysegcnn/core/models.html#UNet.__init__"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#pysegcnn.core.models.UNet.__init__" title="Permalink to this definition">¶</a></dt>
-<dd><p>Initializes internal Module state, shared by both nn.Module and ScriptModule.</p>
+<dd><p>Initialize.</p>
+<dl class="field-list">
+<dt class="field-odd">Parameters</dt>
+<dd class="field-odd"><dl>
+<dt><strong>in_channels</strong><span class="classifier"><cite>int</cite></span></dt><dd><p>Number of channels of the input images.</p>
+</dd>
+<dt><strong>nclasses</strong><span class="classifier"><cite>int</cite></span></dt><dd><p>Number of classes.</p>
+</dd>
+<dt><strong>filters</strong><span class="classifier"><cite>list</cite> [<cite>int</cite>]</span></dt><dd><p>List of input channels to each convolutional block.</p>
+</dd>
+<dt><strong>skip</strong><span class="classifier"><cite>bool</cite></span></dt><dd><p>Whether to apply skip connections from the encoder to the decoder.</p>
+</dd>
+<dt><strong>**kwargs: `dict` [`str`]</strong></dt><dd><p>Additional keyword arguments passed to
+<a class="reference internal" href="pysegcnn.core.layers.Conv2dSame.html#pysegcnn.core.layers.Conv2dSame" title="pysegcnn.core.layers.Conv2dSame"><code class="xref py py-class docutils literal notranslate"><span class="pre">pysegcnn.core.layers.Conv2dSame</span></code></a>.</p>
+</dd>
+</dl>
+</dd>
+</dl>
</dd></dl>
<p class="rubric">Methods</p>
@@ -328,7 +234,7 @@
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#pysegcnn.core.models.UNet.__init__" title="pysegcnn.core.models.UNet.__init__"><code class="xref py py-obj docutils literal notranslate"><span class="pre">__init__</span></code></a>(in_channels, nclasses, filters, …)</p></td>
-<td><p>Initializes internal Module state, shared by both nn.Module and ScriptModule.</p></td>
+<td><p>Initialize.</p></td>
</tr>
<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">add_module</span></code>(name, module)</p></td>
<td><p>Adds a child module to the current module.</p></td>
@@ -468,6 +374,13 @@
</div>
<footer>
+ <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
+
+
+ <a href="pysegcnn.core.models.Network.html" class="btn btn-neutral float-left" title="pysegcnn.core.models.Network" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left"></span> Previous</a>
+
+ </div>
+
<hr/>
diff --git a/docs/_build/html/source/generated/pysegcnn.core.models.html b/docs/_build/html/source/generated/pysegcnn.core.models.html
deleted file mode 100644
index 0c44e06266cbcbee8bbd512adf0045d4685ae8b1..0000000000000000000000000000000000000000
--- a/docs/_build/html/source/generated/pysegcnn.core.models.html
+++ /dev/null
@@ -1,256 +0,0 @@
-
-
-<!DOCTYPE html>
-<html class="writer-html5" lang="en" >
-<head>
- <meta charset="utf-8">
-
- <meta name="viewport" content="width=device-width, initial-scale=1.0">
-
- <title>pysegcnn.core.models — PySegCNN documentation</title>
-
-
-
- <link rel="stylesheet" href="../../_static/css/theme.css" type="text/css" />
- <link rel="stylesheet" href="../../_static/pygments.css" type="text/css" />
-
-
-
-
-
-
-
- <!--[if lt IE 9]>
- <script src="../../_static/js/html5shiv.min.js"></script>
- <![endif]-->
-
-
- <script type="text/javascript" id="documentation_options" data-url_root="../../" src="../../_static/documentation_options.js"></script>
- <script src="../../_static/jquery.js"></script>
- <script src="../../_static/underscore.js"></script>
- <script src="../../_static/doctools.js"></script>
- <script src="../../_static/language_data.js"></script>
-
- <script type="text/javascript" src="../../_static/js/theme.js"></script>
-
-
- <link rel="index" title="Index" href="../../genindex.html" />
- <link rel="search" title="Search" href="../../search.html" />
- <link rel="prev" title="pysegcnn.core.layers" href="pysegcnn.core.layers.html" />
-</head>
-
-<body class="wy-body-for-nav">
-
-
- <div class="wy-grid-for-nav">
-
- <nav data-toggle="wy-nav-shift" class="wy-nav-side">
- <div class="wy-side-scroll">
- <div class="wy-side-nav-search" >
-
-
-
- <a href="../../index.html" class="icon icon-home" alt="Documentation Home"> PySegCNN
-
-
-
- </a>
-
-
-
-
-
-
-
-<div role="search">
- <form id="rtd-search-form" class="wy-form" action="../../search.html" method="get">
- <input type="text" name="q" placeholder="Search docs" />
- <input type="hidden" name="check_keywords" value="yes" />
- <input type="hidden" name="area" value="default" />
- </form>
-</div>
-
-
- </div>
-
-
- <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
-
-
-
-
-
-
- <ul class="current">
-<li class="toctree-l1"><a class="reference internal" href="../requirements.html">Requirements</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
-<li class="toctree-l1 current"><a class="reference internal" href="../api.html">API Reference</a><ul class="current">
-<li class="toctree-l2 current"><a class="reference internal" href="../api.html#layers-and-models">Layers and Models</a><ul class="current">
-<li class="toctree-l3"><a class="reference internal" href="pysegcnn.core.layers.html">pysegcnn.core.layers</a></li>
-<li class="toctree-l3 current"><a class="current reference internal" href="#">pysegcnn.core.models</a><ul>
-<li class="toctree-l4"><a class="reference internal" href="#license">License</a></li>
-</ul>
-</li>
-</ul>
-</li>
-</ul>
-</li>
-</ul>
-
-
-
- </div>
-
- </div>
- </nav>
-
- <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
-
-
- <nav class="wy-nav-top" aria-label="top navigation">
-
- <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
- <a href="../../index.html">PySegCNN</a>
-
- </nav>
-
-
- <div class="wy-nav-content">
-
- <div class="rst-content">
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-<div role="navigation" aria-label="breadcrumbs navigation">
-
- <ul class="wy-breadcrumbs">
-
- <li><a href="../../index.html" class="icon icon-home"></a> »</li>
-
- <li><a href="../api.html">API Reference</a> »</li>
-
- <li>pysegcnn.core.models</li>
-
-
- <li class="wy-breadcrumbs-aside">
-
-
- <a href="../../_sources/source/generated/pysegcnn.core.models.rst.txt" rel="nofollow"> View page source</a>
-
-
- </li>
-
- </ul>
-
-
- <hr/>
-</div>
- <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
- <div itemprop="articleBody">
-
- <div class="section" id="module-pysegcnn.core.models">
-<span id="pysegcnn-core-models"></span><h1>pysegcnn.core.models<a class="headerlink" href="#module-pysegcnn.core.models" title="Permalink to this headline">¶</a></h1>
-<p>Neural networks for semantic image segmentation.</p>
-<div class="section" id="license">
-<h2>License<a class="headerlink" href="#license" title="Permalink to this headline">¶</a></h2>
-<blockquote>
-<div><p>Copyright (c) 2020 Daniel Frisinghelli</p>
-<p>This source code is licensed under the GNU General Public License v3.</p>
-<p>See the LICENSE file in the repository’s root directory.</p>
-</div></blockquote>
-<p class="rubric">Classes</p>
-<table class="longtable docutils align-default">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">Network</span></code>()</p></td>
-<td><p>Generic Network class.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">SupportedLossFunctions</span></code>(value)</p></td>
-<td><p>Names and corresponding classes of the tested loss functions.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">SupportedModels</span></code>(value)</p></td>
-<td><p>Names and corresponding classes of the implemented models.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">SupportedOptimizers</span></code>(value)</p></td>
-<td><p>Names and corresponding classes of the tested optimizers.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">UNet</span></code>(in_channels, nclasses, filters, skip, …)</p></td>
-<td><p>A PyTorch implementation of <a href="#id1"><span class="problematic" id="id2">`U-Net`_</span></a>.</p></td>
-</tr>
-</tbody>
-</table>
-</div>
-</div>
-
-
- </div>
-
- </div>
- <footer>
-
- <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
-
-
- <a href="pysegcnn.core.layers.html" class="btn btn-neutral float-left" title="pysegcnn.core.layers" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left"></span> Previous</a>
-
- </div>
-
-
- <hr/>
-
- <div role="contentinfo">
- <p>
-
- © Copyright 2020, Daniel Frisinghelli
-
- </p>
- </div>
-
-
-
- Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a
-
- <a href="https://github.com/rtfd/sphinx_rtd_theme">theme</a>
-
- provided by <a href="https://readthedocs.org">Read the Docs</a>.
-
-</footer>
-
- </div>
- </div>
-
- </section>
-
- </div>
-
-
- <script type="text/javascript">
- jQuery(function () {
- SphinxRtdTheme.Navigation.enable(true);
- });
- </script>
-
-
-
-
-
-
-</body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/source/install.html b/docs/_build/html/source/install.html
deleted file mode 100644
index 51ec939aa1bdad0857ae4c7ba63521003d8da231..0000000000000000000000000000000000000000
--- a/docs/_build/html/source/install.html
+++ /dev/null
@@ -1,278 +0,0 @@
-
-
-<!DOCTYPE html>
-<html class="writer-html5" lang="en" >
-<head>
- <meta charset="utf-8">
-
- <meta name="viewport" content="width=device-width, initial-scale=1.0">
-
- <title>Installation — PySegCNN documentation</title>
-
-
-
- <link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
- <link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
-
-
-
-
-
-
-
- <!--[if lt IE 9]>
- <script src="../_static/js/html5shiv.min.js"></script>
- <![endif]-->
-
-
- <script type="text/javascript" id="documentation_options" data-url_root="../" src="../_static/documentation_options.js"></script>
- <script src="../_static/jquery.js"></script>
- <script src="../_static/underscore.js"></script>
- <script src="../_static/doctools.js"></script>
- <script src="../_static/language_data.js"></script>
-
- <script type="text/javascript" src="../_static/js/theme.js"></script>
-
-
- <link rel="index" title="Index" href="../genindex.html" />
- <link rel="search" title="Search" href="../search.html" />
- <link rel="prev" title="PySegCNN: Image segmentation with convolutional neural networks in Python" href="../index.html" />
-</head>
-
-<body class="wy-body-for-nav">
-
-
- <div class="wy-grid-for-nav">
-
- <nav data-toggle="wy-nav-shift" class="wy-nav-side">
- <div class="wy-side-scroll">
- <div class="wy-side-nav-search" >
-
-
-
- <a href="../index.html" class="icon icon-home" alt="Documentation Home"> PySegCNN
-
-
-
- </a>
-
-
-
-
-
-
-
-<div role="search">
- <form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
- <input type="text" name="q" placeholder="Search docs" />
- <input type="hidden" name="check_keywords" value="yes" />
- <input type="hidden" name="area" value="default" />
- </form>
-</div>
-
-
- </div>
-
-
- <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
-
-
-
-
-
-
- <ul class="current">
-<li class="toctree-l1 current"><a class="current reference internal" href="#">Installation</a><ul>
-<li class="toctree-l2"><a class="reference internal" href="#download">Download</a></li>
-<li class="toctree-l2"><a class="reference internal" href="#conda">Conda</a></li>
-<li class="toctree-l2"><a class="reference internal" href="#conda-environments">Conda environments</a></li>
-<li class="toctree-l2"><a class="reference internal" href="#install-pytorch">Install PyTorch</a></li>
-<li class="toctree-l2"><a class="reference internal" href="#install-pysegcnn">Install PySegCNN</a></li>
-</ul>
-</li>
-</ul>
-
-
-
- </div>
-
- </div>
- </nav>
-
- <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
-
-
- <nav class="wy-nav-top" aria-label="top navigation">
-
- <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
- <a href="../index.html">PySegCNN</a>
-
- </nav>
-
-
- <div class="wy-nav-content">
-
- <div class="rst-content">
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-<div role="navigation" aria-label="breadcrumbs navigation">
-
- <ul class="wy-breadcrumbs">
-
- <li><a href="../index.html" class="icon icon-home"></a> »</li>
-
- <li>Installation</li>
-
-
- <li class="wy-breadcrumbs-aside">
-
-
- <a href="../_sources/source/install.rst.txt" rel="nofollow"> View page source</a>
-
-
- </li>
-
- </ul>
-
-
- <hr/>
-</div>
- <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
- <div itemprop="articleBody">
-
- <div class="section" id="installation">
-<h1>Installation<a class="headerlink" href="#installation" title="Permalink to this headline">¶</a></h1>
-<div class="section" id="download">
-<h2>Download<a class="headerlink" href="#download" title="Permalink to this headline">¶</a></h2>
-<p>You can download <code class="docutils literal notranslate"><span class="pre">pysegcnn</span></code> from this repository’s
-<a class="reference external" href="https://gitlab.inf.unibz.it/REMSEN/ccisnow/pysegcnn">website</a>
-or alternatively use <code class="docutils literal notranslate"><span class="pre">git</span></code> from terminal:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>git clone https://gitlab.inf.unibz.it/REMSEN/ccisnow/pysegcnn.git
-</pre></div>
-</div>
-<p>This creates a copy of the repository in your current directory on the file
-system.</p>
-</div>
-<div class="section" id="conda">
-<h2>Conda<a class="headerlink" href="#conda" title="Permalink to this headline">¶</a></h2>
-<p>To install <code class="docutils literal notranslate"><span class="pre">pysegcnn</span></code>, I recommend to use the <code class="docutils literal notranslate"><span class="pre">conda</span></code> package manager.
-You can download <code class="docutils literal notranslate"><span class="pre">conda</span></code> <a class="reference external" href="https://docs.conda.io/en/latest/miniconda.html">here</a>.
-Once successfully installed <code class="docutils literal notranslate"><span class="pre">conda</span></code>, I recommend to add <code class="docutils literal notranslate"><span class="pre">conda-forge</span></code> as
-your default channel:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>conda config --add channels conda-forge
-</pre></div>
-</div>
-</div>
-<div class="section" id="conda-environments">
-<h2>Conda environments<a class="headerlink" href="#conda-environments" title="Permalink to this headline">¶</a></h2>
-<p>To install <code class="docutils literal notranslate"><span class="pre">pysegcnn</span></code>, I recommend to create a specific <code class="docutils literal notranslate"><span class="pre">conda</span></code>
-<a class="reference external" href="https://docs.conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html">environment</a>,
-by using the provided <code class="docutils literal notranslate"><span class="pre">environment.yml</span></code> file. In a terminal, navigate to the
-<strong>cloned git repositories root directory</strong> (<code class="docutils literal notranslate"><span class="pre">/pysegcnn</span></code>) and type:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>conda env create -f environment.yml
-</pre></div>
-</div>
-<p>This may take a while. The above command creates a conda environment with all
-required dependencies except the <code class="docutils literal notranslate"><span class="pre">pytorch</span></code> package. The first line in
-<code class="docutils literal notranslate"><span class="pre">environment.yml</span></code> defines the environment name, in this case <code class="docutils literal notranslate"><span class="pre">pysegcnn</span></code>.
-Activate your environment using:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>conda activate pysegcnn
-</pre></div>
-</div>
-</div>
-<div class="section" id="install-pytorch">
-<h2>Install PyTorch<a class="headerlink" href="#install-pytorch" title="Permalink to this headline">¶</a></h2>
-<p>The installation of <code class="docutils literal notranslate"><span class="pre">pytorch</span></code> is heavily dependent on the hardware of your
-machine. Therefore, after activating your environment, install the version of
-<a class="reference external" href="https://pytorch.org/">PyTorch</a> that your system supports by following the
-official <a class="reference external" href="https://pytorch.org/get-started/locally/">instructions</a>.</p>
-<p>If you have to build <code class="docutils literal notranslate"><span class="pre">pytorch</span></code> from source, follow this
-<a class="reference external" href="https://github.com/pytorch/pytorch#from-source">guide</a>.</p>
-<p><a class="reference external" href="https://pytorch.org/get-started/locally/#linux-verification">Here</a> is how to
-check if <code class="docutils literal notranslate"><span class="pre">pytorch</span></code> is installed correctly.</p>
-</div>
-<div class="section" id="install-pysegcnn">
-<h2>Install PySegCNN<a class="headerlink" href="#install-pysegcnn" title="Permalink to this headline">¶</a></h2>
-<p>To finally install <code class="docutils literal notranslate"><span class="pre">pysegcnn</span></code> run the below command <strong>from this repositories
-root directory within the activated ``pysegcnn`` conda environment</strong>:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>pip install -e .
-</pre></div>
-</div>
-<p>If successful, you should be able to import <code class="docutils literal notranslate"><span class="pre">pysegcnn</span></code> from any Python
-interpreter using:</p>
-<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="kn">import</span> <span class="nn">pysegcnn</span>
-</pre></div>
-</div>
-</div>
-</div>
-
-
- </div>
-
- </div>
- <footer>
-
- <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
-
-
- <a href="../index.html" class="btn btn-neutral float-left" title="PySegCNN: Image segmentation with convolutional neural networks in Python" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left"></span> Previous</a>
-
- </div>
-
-
- <hr/>
-
- <div role="contentinfo">
- <p>
-
- © Copyright 2020, Daniel Frisinghelli
-
- </p>
- </div>
-
-
-
- Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a
-
- <a href="https://github.com/rtfd/sphinx_rtd_theme">theme</a>
-
- provided by <a href="https://readthedocs.org">Read the Docs</a>.
-
-</footer>
-
- </div>
- </div>
-
- </section>
-
- </div>
-
-
- <script type="text/javascript">
- jQuery(function () {
- SphinxRtdTheme.Navigation.enable(true);
- });
- </script>
-
-
-
-
-
-
-</body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/source/installation.html b/docs/_build/html/source/installation.html
index 5fb964c0a318f45dcc8d1f48e40ea536541c4678..422e882caa3279a88eb1d2a138bb6ddfec8b1289 100644
--- a/docs/_build/html/source/installation.html
+++ b/docs/_build/html/source/installation.html
@@ -36,6 +36,7 @@
<link rel="index" title="Index" href="../genindex.html" />
<link rel="search" title="Search" href="../search.html" />
+ <link rel="next" title="API Reference" href="api.html" />
<link rel="prev" title="Requirements" href="requirements.html" />
</head>
@@ -91,6 +92,7 @@
<li class="toctree-l2"><a class="reference internal" href="#install-pysegcnn">Install PySegCNN</a></li>
</ul>
</li>
+<li class="toctree-l1"><a class="reference internal" href="api.html">API Reference</a></li>
</ul>
@@ -230,6 +232,8 @@ interpreter using:</p>
<div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
+ <a href="api.html" class="btn btn-neutral float-right" title="API Reference" accesskey="n" rel="next">Next <span class="fa fa-arrow-circle-right"></span></a>
+
<a href="requirements.html" class="btn btn-neutral float-left" title="Requirements" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left"></span> Previous</a>
diff --git a/docs/_build/html/source/main.html b/docs/_build/html/source/main.html
deleted file mode 100644
index d25c64c84a9589ba1a65c9dc349ad910cb737b7c..0000000000000000000000000000000000000000
--- a/docs/_build/html/source/main.html
+++ /dev/null
@@ -1,300 +0,0 @@
-
-
-<!DOCTYPE html>
-<html class="writer-html5" lang="en" >
-<head>
- <meta charset="utf-8">
-
- <meta name="viewport" content="width=device-width, initial-scale=1.0">
-
- <title>main package — PySegCNN documentation</title>
-
-
-
- <link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
- <link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
-
-
-
-
-
-
-
- <!--[if lt IE 9]>
- <script src="../_static/js/html5shiv.min.js"></script>
- <![endif]-->
-
-
- <script type="text/javascript" id="documentation_options" data-url_root="../" src="../_static/documentation_options.js"></script>
- <script src="../_static/jquery.js"></script>
- <script src="../_static/underscore.js"></script>
- <script src="../_static/doctools.js"></script>
- <script src="../_static/language_data.js"></script>
-
- <script type="text/javascript" src="../_static/js/theme.js"></script>
-
-
- <link rel="index" title="Index" href="../genindex.html" />
- <link rel="search" title="Search" href="../search.html" />
- <link rel="next" title="preprocessing package" href="preprocessing.html" />
- <link rel="prev" title="core package" href="core.html" />
-</head>
-
-<body class="wy-body-for-nav">
-
-
- <div class="wy-grid-for-nav">
-
- <nav data-toggle="wy-nav-shift" class="wy-nav-side">
- <div class="wy-side-scroll">
- <div class="wy-side-nav-search" >
-
-
-
- <a href="../index.html" class="icon icon-home" alt="Documentation Home"> PySegCNN
-
-
-
- </a>
-
-
-
-
-
-
-
-<div role="search">
- <form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
- <input type="text" name="q" placeholder="Search docs" />
- <input type="hidden" name="check_keywords" value="yes" />
- <input type="hidden" name="area" value="default" />
- </form>
-</div>
-
-
- </div>
-
-
- <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
-
-
-
-
-
-
- <ul class="current">
-<li class="toctree-l1 current"><a class="reference internal" href="modules.html">pysegcnn</a><ul class="current">
-<li class="toctree-l2"><a class="reference internal" href="core.html">core package</a></li>
-<li class="toctree-l2 current"><a class="current reference internal" href="#">main package</a><ul>
-<li class="toctree-l3"><a class="reference internal" href="#submodules">Submodules</a></li>
-<li class="toctree-l3"><a class="reference internal" href="#module-main.config">main.config module</a><ul>
-<li class="toctree-l4"><a class="reference internal" href="#license">License</a></li>
-</ul>
-</li>
-<li class="toctree-l3"><a class="reference internal" href="#module-main.eval">main.eval module</a><ul>
-<li class="toctree-l4"><a class="reference internal" href="#id1">License</a></li>
-</ul>
-</li>
-<li class="toctree-l3"><a class="reference internal" href="#module-main.train">main.train module</a><ul>
-<li class="toctree-l4"><a class="reference internal" href="#id2">License</a></li>
-</ul>
-</li>
-<li class="toctree-l3"><a class="reference internal" href="#module-main">Module contents</a></li>
-</ul>
-</li>
-<li class="toctree-l2"><a class="reference internal" href="preprocessing.html">preprocessing package</a></li>
-</ul>
-</li>
-</ul>
-
-
-
- </div>
-
- </div>
- </nav>
-
- <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
-
-
- <nav class="wy-nav-top" aria-label="top navigation">
-
- <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
- <a href="../index.html">PySegCNN</a>
-
- </nav>
-
-
- <div class="wy-nav-content">
-
- <div class="rst-content">
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-<div role="navigation" aria-label="breadcrumbs navigation">
-
- <ul class="wy-breadcrumbs">
-
- <li><a href="../index.html" class="icon icon-home"></a> »</li>
-
- <li><a href="modules.html">pysegcnn</a> »</li>
-
- <li>main package</li>
-
-
- <li class="wy-breadcrumbs-aside">
-
-
- <a href="../_sources/source/main.rst.txt" rel="nofollow"> View page source</a>
-
-
- </li>
-
- </ul>
-
-
- <hr/>
-</div>
- <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
- <div itemprop="articleBody">
-
- <div class="section" id="main-package">
-<h1>main package<a class="headerlink" href="#main-package" title="Permalink to this headline">¶</a></h1>
-<div class="section" id="submodules">
-<h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this headline">¶</a></h2>
-</div>
-<div class="section" id="module-main.config">
-<span id="main-config-module"></span><h2>main.config module<a class="headerlink" href="#module-main.config" title="Permalink to this headline">¶</a></h2>
-<p>The configuration file to train and evaluate a model.</p>
-<p>The configuration is handled by the configuration dictionaries.</p>
-<p>Modify the values to your needs, but DO NOT modify the keys.</p>
-<div class="section" id="license">
-<h3>License<a class="headerlink" href="#license" title="Permalink to this headline">¶</a></h3>
-<blockquote>
-<div><p>Copyright (c) 2020 Daniel Frisinghelli</p>
-<p>This source code is licensed under the GNU General Public License v3.</p>
-<p>See the LICENSE file in the repository’s root directory.</p>
-</div></blockquote>
-</div>
-</div>
-<div class="section" id="module-main.eval">
-<span id="main-eval-module"></span><h2>main.eval module<a class="headerlink" href="#module-main.eval" title="Permalink to this headline">¶</a></h2>
-<p>Main script to evaluate a model.</p>
-<p>Steps to run a model evaluation:</p>
-<blockquote>
-<div><ol class="arabic simple">
-<li><p>Configure the dictionary ‘eval_config’ in pysegcnn/main/config.py</p></li>
-<li><p>Save pysegcnn/main/config.py</p></li>
-<li><p>In a terminal, navigate to the repository’s root directory</p></li>
-<li><p>run “python pysegcnn/main/eval.pyâ€</p></li>
-</ol>
-</div></blockquote>
-<div class="section" id="id1">
-<h3>License<a class="headerlink" href="#id1" title="Permalink to this headline">¶</a></h3>
-<blockquote>
-<div><p>Copyright (c) 2020 Daniel Frisinghelli</p>
-<p>This source code is licensed under the GNU General Public License v3.</p>
-<p>See the LICENSE file in the repository’s root directory.</p>
-</div></blockquote>
-</div>
-</div>
-<div class="section" id="module-main.train">
-<span id="main-train-module"></span><h2>main.train module<a class="headerlink" href="#module-main.train" title="Permalink to this headline">¶</a></h2>
-<p>Main script to train a model.</p>
-<p>Steps to launch a model run:</p>
-<blockquote>
-<div><ol class="arabic simple">
-<li><p>Configure the model run in pysegcnn/main/config.py
-(i) configure the dataset : dictionary ‘dataset_config’
-(j) configure the dataset split: dictionary ‘split_config’
-(k) configure the model : dictionary ‘model_config’</p></li>
-<li><p>Save pysegcnn/main/config.py</p></li>
-<li><p>In a terminal, navigate to the repository’s root directory</p></li>
-<li><p>run “python pysegcnn/main/train.pyâ€</p></li>
-</ol>
-</div></blockquote>
-<div class="section" id="id2">
-<h3>License<a class="headerlink" href="#id2" title="Permalink to this headline">¶</a></h3>
-<blockquote>
-<div><p>Copyright (c) 2020 Daniel Frisinghelli</p>
-<p>This source code is licensed under the GNU General Public License v3.</p>
-<p>See the LICENSE file in the repository’s root directory.</p>
-</div></blockquote>
-</div>
-</div>
-<div class="section" id="module-main">
-<span id="module-contents"></span><h2>Module contents<a class="headerlink" href="#module-main" title="Permalink to this headline">¶</a></h2>
-</div>
-</div>
-
-
- </div>
-
- </div>
- <footer>
-
- <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
-
- <a href="preprocessing.html" class="btn btn-neutral float-right" title="preprocessing package" accesskey="n" rel="next">Next <span class="fa fa-arrow-circle-right"></span></a>
-
-
- <a href="core.html" class="btn btn-neutral float-left" title="core package" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left"></span> Previous</a>
-
- </div>
-
-
- <hr/>
-
- <div role="contentinfo">
- <p>
-
- © Copyright 2020, Daniel Frisinghelli
-
- </p>
- </div>
-
-
-
- Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a
-
- <a href="https://github.com/rtfd/sphinx_rtd_theme">theme</a>
-
- provided by <a href="https://readthedocs.org">Read the Docs</a>.
-
-</footer>
-
- </div>
- </div>
-
- </section>
-
- </div>
-
-
- <script type="text/javascript">
- jQuery(function () {
- SphinxRtdTheme.Navigation.enable(true);
- });
- </script>
-
-
-
-
-
-
-</body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/source/modules.html b/docs/_build/html/source/modules.html
deleted file mode 100644
index abe9a8d8ae74a8997a507131c881b262dae29b43..0000000000000000000000000000000000000000
--- a/docs/_build/html/source/modules.html
+++ /dev/null
@@ -1,300 +0,0 @@
-
-
-<!DOCTYPE html>
-<html class="writer-html5" lang="en" >
-<head>
- <meta charset="utf-8">
-
- <meta name="viewport" content="width=device-width, initial-scale=1.0">
-
- <title>pysegcnn — PySegCNN documentation</title>
-
-
-
- <link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
- <link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
-
-
-
-
-
-
-
- <!--[if lt IE 9]>
- <script src="../_static/js/html5shiv.min.js"></script>
- <![endif]-->
-
-
- <script type="text/javascript" id="documentation_options" data-url_root="../" src="../_static/documentation_options.js"></script>
- <script src="../_static/jquery.js"></script>
- <script src="../_static/underscore.js"></script>
- <script src="../_static/doctools.js"></script>
- <script src="../_static/language_data.js"></script>
-
- <script type="text/javascript" src="../_static/js/theme.js"></script>
-
-
- <link rel="index" title="Index" href="../genindex.html" />
- <link rel="search" title="Search" href="../search.html" />
- <link rel="next" title="core package" href="core.html" />
- <link rel="prev" title="<no title>" href="../index.html" />
-</head>
-
-<body class="wy-body-for-nav">
-
-
- <div class="wy-grid-for-nav">
-
- <nav data-toggle="wy-nav-shift" class="wy-nav-side">
- <div class="wy-side-scroll">
- <div class="wy-side-nav-search" >
-
-
-
- <a href="../index.html" class="icon icon-home" alt="Documentation Home"> PySegCNN
-
-
-
- </a>
-
-
-
-
-
-
-
-<div role="search">
- <form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
- <input type="text" name="q" placeholder="Search docs" />
- <input type="hidden" name="check_keywords" value="yes" />
- <input type="hidden" name="area" value="default" />
- </form>
-</div>
-
-
- </div>
-
-
- <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
-
-
-
-
-
-
- <ul class="current">
-<li class="toctree-l1 current"><a class="current reference internal" href="#">pysegcnn</a><ul>
-<li class="toctree-l2"><a class="reference internal" href="core.html">core package</a></li>
-<li class="toctree-l2"><a class="reference internal" href="main.html">main package</a></li>
-<li class="toctree-l2"><a class="reference internal" href="preprocessing.html">preprocessing package</a></li>
-</ul>
-</li>
-</ul>
-
-
-
- </div>
-
- </div>
- </nav>
-
- <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
-
-
- <nav class="wy-nav-top" aria-label="top navigation">
-
- <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
- <a href="../index.html">PySegCNN</a>
-
- </nav>
-
-
- <div class="wy-nav-content">
-
- <div class="rst-content">
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-<div role="navigation" aria-label="breadcrumbs navigation">
-
- <ul class="wy-breadcrumbs">
-
- <li><a href="../index.html" class="icon icon-home"></a> »</li>
-
- <li>pysegcnn</li>
-
-
- <li class="wy-breadcrumbs-aside">
-
-
- <a href="../_sources/source/modules.rst.txt" rel="nofollow"> View page source</a>
-
-
- </li>
-
- </ul>
-
-
- <hr/>
-</div>
- <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
- <div itemprop="articleBody">
-
- <div class="section" id="pysegcnn">
-<h1>pysegcnn<a class="headerlink" href="#pysegcnn" title="Permalink to this headline">¶</a></h1>
-<div class="toctree-wrapper compound">
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="core.html">core package</a><ul>
-<li class="toctree-l2"><a class="reference internal" href="core.html#submodules">Submodules</a></li>
-<li class="toctree-l2"><a class="reference internal" href="core.html#module-core.cli">core.cli module</a><ul>
-<li class="toctree-l3"><a class="reference internal" href="core.html#license">License</a></li>
-</ul>
-</li>
-<li class="toctree-l2"><a class="reference internal" href="core.html#module-core.constants">core.constants module</a><ul>
-<li class="toctree-l3"><a class="reference internal" href="core.html#id1">License</a></li>
-</ul>
-</li>
-<li class="toctree-l2"><a class="reference internal" href="core.html#module-core.dataset">core.dataset module</a><ul>
-<li class="toctree-l3"><a class="reference internal" href="core.html#id2">License</a></li>
-</ul>
-</li>
-<li class="toctree-l2"><a class="reference internal" href="core.html#module-core.graphics">core.graphics module</a><ul>
-<li class="toctree-l3"><a class="reference internal" href="core.html#id4">License</a></li>
-</ul>
-</li>
-<li class="toctree-l2"><a class="reference internal" href="core.html#module-core.layers">core.layers module</a><ul>
-<li class="toctree-l3"><a class="reference internal" href="core.html#id5">License</a></li>
-</ul>
-</li>
-<li class="toctree-l2"><a class="reference internal" href="core.html#module-core.logging">core.logging module</a><ul>
-<li class="toctree-l3"><a class="reference internal" href="core.html#id6">License</a></li>
-</ul>
-</li>
-<li class="toctree-l2"><a class="reference internal" href="core.html#module-core.models">core.models module</a><ul>
-<li class="toctree-l3"><a class="reference internal" href="core.html#id7">License</a></li>
-</ul>
-</li>
-<li class="toctree-l2"><a class="reference internal" href="core.html#module-core.predict">core.predict module</a><ul>
-<li class="toctree-l3"><a class="reference internal" href="core.html#id8">License</a></li>
-</ul>
-</li>
-<li class="toctree-l2"><a class="reference internal" href="core.html#module-core.split">core.split module</a><ul>
-<li class="toctree-l3"><a class="reference internal" href="core.html#id9">License</a></li>
-</ul>
-</li>
-<li class="toctree-l2"><a class="reference internal" href="core.html#module-core.trainer">core.trainer module</a><ul>
-<li class="toctree-l3"><a class="reference internal" href="core.html#id10">License</a></li>
-</ul>
-</li>
-<li class="toctree-l2"><a class="reference internal" href="core.html#module-core.transforms">core.transforms module</a><ul>
-<li class="toctree-l3"><a class="reference internal" href="core.html#id11">License</a></li>
-</ul>
-</li>
-<li class="toctree-l2"><a class="reference internal" href="core.html#module-core.utils">core.utils module</a><ul>
-<li class="toctree-l3"><a class="reference internal" href="core.html#id12">License</a></li>
-</ul>
-</li>
-<li class="toctree-l2"><a class="reference internal" href="core.html#module-core">Module contents</a></li>
-</ul>
-</li>
-<li class="toctree-l1"><a class="reference internal" href="main.html">main package</a><ul>
-<li class="toctree-l2"><a class="reference internal" href="main.html#submodules">Submodules</a></li>
-<li class="toctree-l2"><a class="reference internal" href="main.html#module-main.config">main.config module</a><ul>
-<li class="toctree-l3"><a class="reference internal" href="main.html#license">License</a></li>
-</ul>
-</li>
-<li class="toctree-l2"><a class="reference internal" href="main.html#module-main.eval">main.eval module</a><ul>
-<li class="toctree-l3"><a class="reference internal" href="main.html#id1">License</a></li>
-</ul>
-</li>
-<li class="toctree-l2"><a class="reference internal" href="main.html#module-main.train">main.train module</a><ul>
-<li class="toctree-l3"><a class="reference internal" href="main.html#id2">License</a></li>
-</ul>
-</li>
-<li class="toctree-l2"><a class="reference internal" href="main.html#module-main">Module contents</a></li>
-</ul>
-</li>
-<li class="toctree-l1"><a class="reference internal" href="preprocessing.html">preprocessing package</a><ul>
-<li class="toctree-l2"><a class="reference internal" href="preprocessing.html#submodules">Submodules</a></li>
-<li class="toctree-l2"><a class="reference internal" href="preprocessing.html#module-preprocessing.sparcs">preprocessing.sparcs module</a><ul>
-<li class="toctree-l3"><a class="reference internal" href="preprocessing.html#license">License</a></li>
-</ul>
-</li>
-<li class="toctree-l2"><a class="reference internal" href="preprocessing.html#module-preprocessing">Module contents</a></li>
-</ul>
-</li>
-</ul>
-</div>
-</div>
-
-
- </div>
-
- </div>
- <footer>
-
- <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
-
- <a href="core.html" class="btn btn-neutral float-right" title="core package" accesskey="n" rel="next">Next <span class="fa fa-arrow-circle-right"></span></a>
-
-
- <a href="../index.html" class="btn btn-neutral float-left" title="<no title>" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left"></span> Previous</a>
-
- </div>
-
-
- <hr/>
-
- <div role="contentinfo">
- <p>
-
- © Copyright 2020, Daniel Frisinghelli
-
- </p>
- </div>
-
-
-
- Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a
-
- <a href="https://github.com/rtfd/sphinx_rtd_theme">theme</a>
-
- provided by <a href="https://readthedocs.org">Read the Docs</a>.
-
-</footer>
-
- </div>
- </div>
-
- </section>
-
- </div>
-
-
- <script type="text/javascript">
- jQuery(function () {
- SphinxRtdTheme.Navigation.enable(true);
- });
- </script>
-
-
-
-
-
-
-</body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/source/preprocessing.html b/docs/_build/html/source/preprocessing.html
deleted file mode 100644
index 9621025cb91bbd53068b23cbc577d6e7ff9953d6..0000000000000000000000000000000000000000
--- a/docs/_build/html/source/preprocessing.html
+++ /dev/null
@@ -1,247 +0,0 @@
-
-
-<!DOCTYPE html>
-<html class="writer-html5" lang="en" >
-<head>
- <meta charset="utf-8">
-
- <meta name="viewport" content="width=device-width, initial-scale=1.0">
-
- <title>preprocessing package — PySegCNN documentation</title>
-
-
-
- <link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
- <link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
-
-
-
-
-
-
-
- <!--[if lt IE 9]>
- <script src="../_static/js/html5shiv.min.js"></script>
- <![endif]-->
-
-
- <script type="text/javascript" id="documentation_options" data-url_root="../" src="../_static/documentation_options.js"></script>
- <script src="../_static/jquery.js"></script>
- <script src="../_static/underscore.js"></script>
- <script src="../_static/doctools.js"></script>
- <script src="../_static/language_data.js"></script>
-
- <script type="text/javascript" src="../_static/js/theme.js"></script>
-
-
- <link rel="index" title="Index" href="../genindex.html" />
- <link rel="search" title="Search" href="../search.html" />
- <link rel="prev" title="main package" href="main.html" />
-</head>
-
-<body class="wy-body-for-nav">
-
-
- <div class="wy-grid-for-nav">
-
- <nav data-toggle="wy-nav-shift" class="wy-nav-side">
- <div class="wy-side-scroll">
- <div class="wy-side-nav-search" >
-
-
-
- <a href="../index.html" class="icon icon-home" alt="Documentation Home"> PySegCNN
-
-
-
- </a>
-
-
-
-
-
-
-
-<div role="search">
- <form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
- <input type="text" name="q" placeholder="Search docs" />
- <input type="hidden" name="check_keywords" value="yes" />
- <input type="hidden" name="area" value="default" />
- </form>
-</div>
-
-
- </div>
-
-
- <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
-
-
-
-
-
-
- <ul class="current">
-<li class="toctree-l1 current"><a class="reference internal" href="modules.html">pysegcnn</a><ul class="current">
-<li class="toctree-l2"><a class="reference internal" href="core.html">core package</a></li>
-<li class="toctree-l2"><a class="reference internal" href="main.html">main package</a></li>
-<li class="toctree-l2 current"><a class="current reference internal" href="#">preprocessing package</a><ul>
-<li class="toctree-l3"><a class="reference internal" href="#submodules">Submodules</a></li>
-<li class="toctree-l3"><a class="reference internal" href="#module-preprocessing.sparcs">preprocessing.sparcs module</a><ul>
-<li class="toctree-l4"><a class="reference internal" href="#license">License</a></li>
-</ul>
-</li>
-<li class="toctree-l3"><a class="reference internal" href="#module-preprocessing">Module contents</a></li>
-</ul>
-</li>
-</ul>
-</li>
-</ul>
-
-
-
- </div>
-
- </div>
- </nav>
-
- <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
-
-
- <nav class="wy-nav-top" aria-label="top navigation">
-
- <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
- <a href="../index.html">PySegCNN</a>
-
- </nav>
-
-
- <div class="wy-nav-content">
-
- <div class="rst-content">
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-<div role="navigation" aria-label="breadcrumbs navigation">
-
- <ul class="wy-breadcrumbs">
-
- <li><a href="../index.html" class="icon icon-home"></a> »</li>
-
- <li><a href="modules.html">pysegcnn</a> »</li>
-
- <li>preprocessing package</li>
-
-
- <li class="wy-breadcrumbs-aside">
-
-
- <a href="../_sources/source/preprocessing.rst.txt" rel="nofollow"> View page source</a>
-
-
- </li>
-
- </ul>
-
-
- <hr/>
-</div>
- <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
- <div itemprop="articleBody">
-
- <div class="section" id="preprocessing-package">
-<h1>preprocessing package<a class="headerlink" href="#preprocessing-package" title="Permalink to this headline">¶</a></h1>
-<div class="section" id="submodules">
-<h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this headline">¶</a></h2>
-</div>
-<div class="section" id="module-preprocessing.sparcs">
-<span id="preprocessing-sparcs-module"></span><h2>preprocessing.sparcs module<a class="headerlink" href="#module-preprocessing.sparcs" title="Permalink to this headline">¶</a></h2>
-<p>Command line script to preprocess the Sparcs dataset.</p>
-<p>After downloading the Sparcs dataset, from the repository’s root directory run:</p>
-<blockquote>
-<div><p>python pysegcnn/preprocessing/sparcs.py</p>
-</div></blockquote>
-<p>The above command will print the usage of the script.</p>
-<div class="section" id="license">
-<h3>License<a class="headerlink" href="#license" title="Permalink to this headline">¶</a></h3>
-<blockquote>
-<div><p>Copyright (c) 2020 Daniel Frisinghelli</p>
-<p>This source code is licensed under the GNU General Public License v3.</p>
-<p>See the LICENSE file in the repository’s root directory.</p>
-</div></blockquote>
-</div>
-</div>
-<div class="section" id="module-preprocessing">
-<span id="module-contents"></span><h2>Module contents<a class="headerlink" href="#module-preprocessing" title="Permalink to this headline">¶</a></h2>
-</div>
-</div>
-
-
- </div>
-
- </div>
- <footer>
-
- <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
-
-
- <a href="main.html" class="btn btn-neutral float-left" title="main package" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left"></span> Previous</a>
-
- </div>
-
-
- <hr/>
-
- <div role="contentinfo">
- <p>
-
- © Copyright 2020, Daniel Frisinghelli
-
- </p>
- </div>
-
-
-
- Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a
-
- <a href="https://github.com/rtfd/sphinx_rtd_theme">theme</a>
-
- provided by <a href="https://readthedocs.org">Read the Docs</a>.
-
-</footer>
-
- </div>
- </div>
-
- </section>
-
- </div>
-
-
- <script type="text/javascript">
- jQuery(function () {
- SphinxRtdTheme.Navigation.enable(true);
- });
- </script>
-
-
-
-
-
-
-</body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/source/readme.html b/docs/_build/html/source/readme.html
deleted file mode 100644
index 32d5d7f0abb36bc9083d5cfceb7b0c6d5d407d9f..0000000000000000000000000000000000000000
--- a/docs/_build/html/source/readme.html
+++ /dev/null
@@ -1,210 +0,0 @@
-
-
-<!DOCTYPE html>
-<html class="writer-html5" lang="en" >
-<head>
- <meta charset="utf-8">
-
- <meta name="viewport" content="width=device-width, initial-scale=1.0">
-
- <title>README — PySegCNN documentation</title>
-
-
-
- <link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
- <link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
-
-
-
-
-
-
-
- <!--[if lt IE 9]>
- <script src="../_static/js/html5shiv.min.js"></script>
- <![endif]-->
-
-
- <script type="text/javascript" id="documentation_options" data-url_root="../" src="../_static/documentation_options.js"></script>
- <script src="../_static/jquery.js"></script>
- <script src="../_static/underscore.js"></script>
- <script src="../_static/doctools.js"></script>
- <script src="../_static/language_data.js"></script>
-
- <script type="text/javascript" src="../_static/js/theme.js"></script>
-
-
- <link rel="index" title="Index" href="../genindex.html" />
- <link rel="search" title="Search" href="../search.html" />
- <link rel="prev" title="preprocessing package" href="preprocessing.html" />
-</head>
-
-<body class="wy-body-for-nav">
-
-
- <div class="wy-grid-for-nav">
-
- <nav data-toggle="wy-nav-shift" class="wy-nav-side">
- <div class="wy-side-scroll">
- <div class="wy-side-nav-search" >
-
-
-
- <a href="../index.html" class="icon icon-home" alt="Documentation Home"> PySegCNN
-
-
-
- </a>
-
-
-
-
-
-
-
-<div role="search">
- <form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
- <input type="text" name="q" placeholder="Search docs" />
- <input type="hidden" name="check_keywords" value="yes" />
- <input type="hidden" name="area" value="default" />
- </form>
-</div>
-
-
- </div>
-
-
- <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
-
-
-
-
-
-
- <ul class="current">
-<li class="toctree-l1"><a class="reference internal" href="modules.html">pysegcnn</a></li>
-<li class="toctree-l1 current"><a class="current reference internal" href="#">README</a></li>
-</ul>
-
-
-
- </div>
-
- </div>
- </nav>
-
- <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
-
-
- <nav class="wy-nav-top" aria-label="top navigation">
-
- <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
- <a href="../index.html">PySegCNN</a>
-
- </nav>
-
-
- <div class="wy-nav-content">
-
- <div class="rst-content">
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-<div role="navigation" aria-label="breadcrumbs navigation">
-
- <ul class="wy-breadcrumbs">
-
- <li><a href="../index.html" class="icon icon-home"></a> »</li>
-
- <li>README</li>
-
-
- <li class="wy-breadcrumbs-aside">
-
-
- <a href="../_sources/source/readme.rst.txt" rel="nofollow"> View page source</a>
-
-
- </li>
-
- </ul>
-
-
- <hr/>
-</div>
- <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
- <div itemprop="articleBody">
-
- <div class="section" id="readme">
-<h1>README<a class="headerlink" href="#readme" title="Permalink to this headline">¶</a></h1>
-</div>
-
-
- </div>
-
- </div>
- <footer>
-
- <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
-
-
- <a href="preprocessing.html" class="btn btn-neutral float-left" title="preprocessing package" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left"></span> Previous</a>
-
- </div>
-
-
- <hr/>
-
- <div role="contentinfo">
- <p>
-
- © Copyright 2020, Daniel Frisinghelli
-
- </p>
- </div>
-
-
-
- Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a
-
- <a href="https://github.com/rtfd/sphinx_rtd_theme">theme</a>
-
- provided by <a href="https://readthedocs.org">Read the Docs</a>.
-
-</footer>
-
- </div>
- </div>
-
- </section>
-
- </div>
-
-
- <script type="text/javascript">
- jQuery(function () {
- SphinxRtdTheme.Navigation.enable(true);
- });
- </script>
-
-
-
-
-
-
-</body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/source/requirements.html b/docs/_build/html/source/requirements.html
index 6360808703ea68989479d2ad31c0f56b73534813..aaceee5fc90ef2aaabacc36c5084a30698421544 100644
--- a/docs/_build/html/source/requirements.html
+++ b/docs/_build/html/source/requirements.html
@@ -92,6 +92,7 @@
</ul>
</li>
<li class="toctree-l1"><a class="reference internal" href="installation.html">Installation</a></li>
+<li class="toctree-l1"><a class="reference internal" href="api.html">API Reference</a></li>
</ul>
diff --git a/docs/conf.py b/docs/conf.py
index 40c453ea0210443b06710df6915585055cda4523..c3ffaec690db1aad2cec9bdc03662c2f87fa6b5f 100644
--- a/docs/conf.py
+++ b/docs/conf.py
@@ -32,15 +32,17 @@ author = 'Daniel Frisinghelli'
# ones.
extensions = [
'sphinx.ext.autodoc',
+ 'sphinx.ext.autosummary',
'sphinx.ext.viewcode',
- 'sphinx.ext.napoleon',
- 'recommonmark',
- 'numpydoc'
+ 'numpydoc',
]
-# napoleon options
-napoleon_google_docstring = False
+# numpydoc options
+numpydoc_show_class_members = False
+numpydoc_show_inherited_class_members = False
+# autosummary options
+autosummary_generate = True
# Add any paths that contain templates here, relative to this directory.
templates_path = ['_templates']
diff --git a/docs/source/api.rst b/docs/source/api.rst
index 4de52de54bd3cdf346bfdf3d35e6bef1ce7643ab..6d2aabc35a3d4889c14b817a6580618244188067 100644
--- a/docs/source/api.rst
+++ b/docs/source/api.rst
@@ -13,9 +13,9 @@ Dataset
Custom dataset classes compliant to the PyTorch `standard <https://pytorch.org/docs/stable/data.html>`_.
-Image Dataset
-^^^^^^^^^^^^^
-Generic class to implement `custom datasets <https://pytorch.org/tutorials/beginner/data_loading_tutorial.html>`_.
+Generic classes
+^^^^^^^^^^^^^^^
+Generic class to implement a `custom dataset <https://pytorch.org/tutorials/beginner/data_loading_tutorial.html>`_.
.. autosummary::
:toctree: generated/
@@ -23,11 +23,21 @@ Generic class to implement `custom datasets <https://pytorch.org/tutorials/begin
core.dataset.ImageDataset
-Supported datasets
-^^^^^^^^^^^^^^^^^^
+Generic class to implement a custom dataset following a standard directory
+structure.
+
+.. autosummary::
+ :toctree: generated/
+ :nosignatures:
+
+ core.dataset.StandardEoDataset
+
-The following open-source spaceborne multispectral image datasets are supported
-out-of-the-box:
+Specific classes
+^^^^^^^^^^^^^^^^
+
+Specific classes for some open-source image datasets. Currently, the following
+spaceborne multispectral image datasets are supported out-of-the-box:
.. autosummary::
:toctree: generated/
@@ -36,7 +46,6 @@ out-of-the-box:
core.dataset.SparcsDataset
core.dataset.Cloud95Dataset
-
Models
------
@@ -81,6 +90,9 @@ Neural Networks
core.models.UNet
+..
+ Links:
+
.. _Sparcs:
https://www.usgs.gov/land-resources/nli/landsat/spatial-procedures-automated-removal-cloud-and-shadow-sparcs-validation
@@ -92,3 +104,12 @@ Neural Networks
.. _U-Net:
https://arxiv.org/abs/1505.04597
+
+.. _Landsat-8:
+ https://www.usgs.gov/land-resources/nli/landsat/landsat-8?qt-science_support_page_related_con=0#qt-science_support_page_related_con
+
+.. _Sentinel-2:
+ https://sentinel.esa.int/web/sentinel/missions/sentinel-2
+
+.. _Hughes & Hayes (2014):
+ https://www.mdpi.com/2072-4292/6/6/4907
diff --git a/docs/source/generated/pysegcnn.core.dataset.Cloud95Dataset.rst b/docs/source/generated/pysegcnn.core.dataset.Cloud95Dataset.rst
index 0b8650621242af8de39fc4bf54f43f371b295a94..a1810656ee6dab330ce851fe7a3320f4396a65d0 100644
--- a/docs/source/generated/pysegcnn.core.dataset.Cloud95Dataset.rst
+++ b/docs/source/generated/pysegcnn.core.dataset.Cloud95Dataset.rst
@@ -1,4 +1,4 @@
-pysegcnn.core.dataset.Cloud95Dataset
+pysegcnn.core.dataset.Cloud95Dataset
====================================
.. currentmodule:: pysegcnn.core.dataset
diff --git a/docs/source/generated/pysegcnn.core.dataset.ImageDataset.rst b/docs/source/generated/pysegcnn.core.dataset.ImageDataset.rst
index c04ed1f96ea2d230023e87a20bebeb31c0fe1050..31c36ff73f998226baf6cf9c48daeaea74404f94 100644
--- a/docs/source/generated/pysegcnn.core.dataset.ImageDataset.rst
+++ b/docs/source/generated/pysegcnn.core.dataset.ImageDataset.rst
@@ -1,4 +1,4 @@
-pysegcnn.core.dataset.ImageDataset
+pysegcnn.core.dataset.ImageDataset
==================================
.. currentmodule:: pysegcnn.core.dataset
diff --git a/docs/source/generated/pysegcnn.core.dataset.SparcsDataset.rst b/docs/source/generated/pysegcnn.core.dataset.SparcsDataset.rst
index d5c9fe46ca1985f87d71a260f0a9168ce193c5c4..cb357b5f003e77960920e99d3fe0ec24ec014b46 100644
--- a/docs/source/generated/pysegcnn.core.dataset.SparcsDataset.rst
+++ b/docs/source/generated/pysegcnn.core.dataset.SparcsDataset.rst
@@ -1,4 +1,4 @@
-pysegcnn.core.dataset.SparcsDataset
+pysegcnn.core.dataset.SparcsDataset
===================================
.. currentmodule:: pysegcnn.core.dataset
diff --git a/docs/source/generated/pysegcnn.core.layers.Block.rst b/docs/source/generated/pysegcnn.core.layers.Block.rst
index 3e960b4c77a531221739831ef70e2348e407fd1c..31875348b6f55bb5513dd3785c91d0cddb1a671e 100644
--- a/docs/source/generated/pysegcnn.core.layers.Block.rst
+++ b/docs/source/generated/pysegcnn.core.layers.Block.rst
@@ -1,4 +1,4 @@
-pysegcnn.core.layers.Block
+pysegcnn.core.layers.Block
==========================
.. currentmodule:: pysegcnn.core.layers
diff --git a/docs/source/generated/pysegcnn.core.layers.Conv2dSame.rst b/docs/source/generated/pysegcnn.core.layers.Conv2dSame.rst
index 7c81180f3dbf6d1b08ea77e608a908975395ddb5..42b472e1ec390a7329cfc9813c4a19efb35fe0ff 100644
--- a/docs/source/generated/pysegcnn.core.layers.Conv2dSame.rst
+++ b/docs/source/generated/pysegcnn.core.layers.Conv2dSame.rst
@@ -1,4 +1,4 @@
-pysegcnn.core.layers.Conv2dSame
+pysegcnn.core.layers.Conv2dSame
===============================
.. currentmodule:: pysegcnn.core.layers
diff --git a/docs/source/generated/pysegcnn.core.layers.ConvBnReluMaxPool.rst b/docs/source/generated/pysegcnn.core.layers.ConvBnReluMaxPool.rst
index 7a99902ee622930c6a3ed940f8f2f7e8af705b9d..6260fe50769352e915c3dc3d26ae9b9310ea7c52 100644
--- a/docs/source/generated/pysegcnn.core.layers.ConvBnReluMaxPool.rst
+++ b/docs/source/generated/pysegcnn.core.layers.ConvBnReluMaxPool.rst
@@ -1,4 +1,4 @@
-pysegcnn.core.layers.ConvBnReluMaxPool
+pysegcnn.core.layers.ConvBnReluMaxPool
======================================
.. currentmodule:: pysegcnn.core.layers
diff --git a/docs/source/generated/pysegcnn.core.layers.ConvBnReluMaxUnpool.rst b/docs/source/generated/pysegcnn.core.layers.ConvBnReluMaxUnpool.rst
index c83252a532e4d66427284193e0ce3bd6233ad52e..386abc1be48bcf7487098b4847c7e6385d849772 100644
--- a/docs/source/generated/pysegcnn.core.layers.ConvBnReluMaxUnpool.rst
+++ b/docs/source/generated/pysegcnn.core.layers.ConvBnReluMaxUnpool.rst
@@ -1,4 +1,4 @@
-pysegcnn.core.layers.ConvBnReluMaxUnpool
+pysegcnn.core.layers.ConvBnReluMaxUnpool
========================================
.. currentmodule:: pysegcnn.core.layers
diff --git a/docs/source/generated/pysegcnn.core.layers.ConvBnReluUpsample.rst b/docs/source/generated/pysegcnn.core.layers.ConvBnReluUpsample.rst
index bac34eecc38456d240d6889d01a7e97b9268dd99..6d4f3e6ac7d591f5aaa1da607b6c5730d099d6d0 100644
--- a/docs/source/generated/pysegcnn.core.layers.ConvBnReluUpsample.rst
+++ b/docs/source/generated/pysegcnn.core.layers.ConvBnReluUpsample.rst
@@ -1,4 +1,4 @@
-pysegcnn.core.layers.ConvBnReluUpsample
+pysegcnn.core.layers.ConvBnReluUpsample
=======================================
.. currentmodule:: pysegcnn.core.layers
diff --git a/docs/source/generated/pysegcnn.core.layers.Decoder.rst b/docs/source/generated/pysegcnn.core.layers.Decoder.rst
index 9ca6527d10479cdf03eb3d4213af41da9c295b4a..a5fcf7fd1a0190303deaee9d603008990cc060f5 100644
--- a/docs/source/generated/pysegcnn.core.layers.Decoder.rst
+++ b/docs/source/generated/pysegcnn.core.layers.Decoder.rst
@@ -1,4 +1,4 @@
-pysegcnn.core.layers.Decoder
+pysegcnn.core.layers.Decoder
============================
.. currentmodule:: pysegcnn.core.layers
diff --git a/docs/source/generated/pysegcnn.core.layers.DecoderBlock.rst b/docs/source/generated/pysegcnn.core.layers.DecoderBlock.rst
index 3acc653c6098b27269e7e171bdc7a9f22ab570e8..d6a87bc7f63568ac82cddf0dc5f4d4653391343c 100644
--- a/docs/source/generated/pysegcnn.core.layers.DecoderBlock.rst
+++ b/docs/source/generated/pysegcnn.core.layers.DecoderBlock.rst
@@ -1,4 +1,4 @@
-pysegcnn.core.layers.DecoderBlock
+pysegcnn.core.layers.DecoderBlock
=================================
.. currentmodule:: pysegcnn.core.layers
diff --git a/docs/source/generated/pysegcnn.core.layers.Encoder.rst b/docs/source/generated/pysegcnn.core.layers.Encoder.rst
index 09aa71ccc1c2f5cf0c5fcf0b84a6e47fde490f4e..fcd00c6bc912c46d3ac641b3d2fdeb5a1fd3295c 100644
--- a/docs/source/generated/pysegcnn.core.layers.Encoder.rst
+++ b/docs/source/generated/pysegcnn.core.layers.Encoder.rst
@@ -1,4 +1,4 @@
-pysegcnn.core.layers.Encoder
+pysegcnn.core.layers.Encoder
============================
.. currentmodule:: pysegcnn.core.layers
diff --git a/docs/source/generated/pysegcnn.core.layers.EncoderBlock.rst b/docs/source/generated/pysegcnn.core.layers.EncoderBlock.rst
index 37385b204208ec53e9a247657998a864ad5a0602..454ce4102139d897e53bbad4c61d61bc7b6aeae6 100644
--- a/docs/source/generated/pysegcnn.core.layers.EncoderBlock.rst
+++ b/docs/source/generated/pysegcnn.core.layers.EncoderBlock.rst
@@ -1,4 +1,4 @@
-pysegcnn.core.layers.EncoderBlock
+pysegcnn.core.layers.EncoderBlock
=================================
.. currentmodule:: pysegcnn.core.layers
diff --git a/docs/source/generated/pysegcnn.core.models.Network.rst b/docs/source/generated/pysegcnn.core.models.Network.rst
index a5b21f07fee5ee799539ac7daf52a68abd48d617..99819a3608b928852226184cbf9d39f50697a252 100644
--- a/docs/source/generated/pysegcnn.core.models.Network.rst
+++ b/docs/source/generated/pysegcnn.core.models.Network.rst
@@ -1,4 +1,4 @@
-pysegcnn.core.models.Network
+pysegcnn.core.models.Network
============================
.. currentmodule:: pysegcnn.core.models
diff --git a/docs/source/generated/pysegcnn.core.models.UNet.rst b/docs/source/generated/pysegcnn.core.models.UNet.rst
index 933d3fcbccf6d0d26a1c0bee26300e0fa0cae620..92cb118b24485a58a9d65aafbd223aafadf1ce59 100644
--- a/docs/source/generated/pysegcnn.core.models.UNet.rst
+++ b/docs/source/generated/pysegcnn.core.models.UNet.rst
@@ -1,4 +1,4 @@
-pysegcnn.core.models.UNet
+pysegcnn.core.models.UNet
=========================
.. currentmodule:: pysegcnn.core.models