Skip to content
Snippets Groups Projects
Commit e835e7fc authored by Frisinghelli Daniel's avatar Frisinghelli Daniel
Browse files

Updating documentation.

parent e6d43c8a
No related branches found
No related tags found
No related merge requests found
......@@ -36,6 +36,8 @@ extensions = [
'sphinx.ext.viewcode',
'sphinx.ext.intersphinx',
'numpydoc',
'IPython.sphinxext.ipython_directive',
'IPython.sphinxext.ipython_console_highlighting'
]
# numpydoc options
......
......@@ -18,7 +18,6 @@ Using PySegCNN
.. toctree::
:maxdepth: 1
./source/requirements
./source/installation
./source/api
......
......@@ -5,31 +5,29 @@
API Reference
=============
This page lists the available functions and classes of ``pysegcnn``.
This page lists functions and classes of the ``pysegcnn`` package, which are
relevant at the API level. If you need to dig deeper into the source files, go
to the git `repository <https://gitlab.inf.unibz.it/REMSEN/ccisnow/pysegcnn>`_.
Dataset
-------
Custom dataset classes compliant to the PyTorch `standard <https://pytorch.org/docs/stable/data.html>`_.
The ``pysegcnn`` package offers support for custom datasets, which are compliant
to the PyTorch `standard <https://pytorch.org/docs/stable/data.html>`_.
Currently, image datasets from the `Landsat <https://landsat.usgs.gov/>`_ and
`Sentinel-2`_ satellites are supported.
Generic classes
^^^^^^^^^^^^^^^
Generic class to implement a `custom dataset <https://pytorch.org/tutorials/beginner/data_loading_tutorial.html>`_.
.. autosummary::
:toctree: generated/
:nosignatures:
core.dataset.ImageDataset
Generic class to implement a custom dataset following a standard directory
structure.
The following two generic classes can be inherited to implement a
`custom dataset <https://pytorch.org/tutorials/beginner/data_loading_tutorial.html>`_.
.. autosummary::
:toctree: generated/
:nosignatures:
core.dataset.ImageDataset
core.dataset.StandardEoDataset
......@@ -49,38 +47,24 @@ spaceborne multispectral image datasets are supported out-of-the-box:
Models
------
Layers
^^^^^^
Convolutional neural network layers.
The ``pysegcnn`` package ships with a customizable interface to build
convolutional neural networks for image segmentation.
.. autosummary::
:toctree: generated/
:nosignatures:
core.layers.Block
core.layers.Conv2dSame
core.layers.Block
core.layers.ConvBnReluMaxPool
core.layers.ConvBnReluMaxUnpool
core.layers.ConvBnReluUpsample
Encoder-Decoder architechture
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Generic ``Encoder`` and ``Decoder`` classes to build an encoder-decoder
architecture.
.. autosummary::
:toctree: generated/
:nosignatures:
core.layers.EncoderBlock
core.layers.DecoderBlock
core.layers.Encoder
core.layers.Decoder
Neural Networks
^^^^^^^^^^^^^^^
Implemented models.
.. autosummary::
:toctree: generated/
......@@ -90,6 +74,19 @@ Neural Networks
core.models.UNet
Training, Validation and Test set
---------------------------------
Classes to split a dataset into training, validation and test set.
.. autosummary::
:toctree: generated/
:nosignatures:
core.split.DateSplit
core.split.RandomTileSplit
core.split.RandomSceneSplit
..
Links:
......@@ -113,3 +110,6 @@ Neural Networks
.. _Hughes & Hayes (2014):
https://www.mdpi.com/2072-4292/6/6/4907
.. _Early Stopping:
https://en.wikipedia.org/wiki/Early_stopping
......@@ -3,6 +3,22 @@
Installation
============
Requirements
------------
``pysegcnn`` is a pure Python package that works on both Windows and Linux.
.. important::
``pysegcnn`` requires ``Python>=3.7``.
Here is a list of all dependencies of ``pysegcnn``.
- numpy
- scipy
- matplotlib
- pytorch
- gdal
Download
---------
You can download ``pysegcnn`` from this repository's
......@@ -74,6 +90,6 @@ root directory within the activated ``pysegcnn`` conda environment**:
If successful, you should be able to import ``pysegcnn`` from any Python
interpreter using:
.. code-block:: python
.. ipython:: python
import pysegcnn
Requirements
============
``pysegcnn`` is a pure Python package that works on both Windows and Linux.
.. important:
``pysegcnn`` requires ``Python >= 3.7``.
Dependencies
^^^^^^^^^^^^
Here is a list of all dependencies of ``pysegcnn``.
Machine learning framework
^^^^^^^^^^^^^^^^^^^^^^^^^^
- pytorch
SciPy stack
^^^^^^^^^^^
- numpy
- scipy
- matplotlib
Raster tools
^^^^^^^^^^^^
- gdal
Testing
^^^^^^^
- pytest
And if you want to build the documentation:
- sphinx
- sphinx_rtd_theme
If you are using `conda <https://docs.conda.io/en/latest/miniconda.html>`_,
you can use the ``environment.yml`` file that ships with ``pysegcnn`` to take
care of the dependencies, see :doc:`installation`.
\ No newline at end of file
0% Loading or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment