Skip to content
GitLab
Explore
Sign in
Primary navigation
Search or go to…
Project
P
PySegCNN
Manage
Activity
Members
Labels
Plan
Issues
Issue boards
Milestones
Wiki
Code
Merge requests
Repository
Branches
Commits
Tags
Repository graph
Compare revisions
Snippets
Deploy
Releases
Package registry
Container Registry
Model registry
Operate
Terraform modules
Monitor
Incidents
Analyze
Value stream analytics
Contributor analytics
Repository analytics
Model experiments
Help
Help
Support
GitLab documentation
Compare GitLab plans
Community forum
Contribute to GitLab
Provide feedback
Keyboard shortcuts
?
Snippets
Groups
Projects
Show more breadcrumbs
earth_observation_public
PySegCNN
Repository graph
Repository graph
You can move around the graph by using the arrow keys.
807f462886b4856e043960830d60d8358c4dff78
Select Git revision
Branches
1
master
default
protected
1 result
Begin with the selected commit
Created with Raphaël 2.2.0
25
Mar
24
23
22
15
12
11
10
9
3
2
26
Feb
25
24
20
19
18
17
16
12
10
9
5
4
3
2
1
29
Jan
28
27
26
25
22
20
19
15
14
13
12
11
8
7
23
Dec
21
17
16
15
14
11
9
4
27
Nov
26
25
9
Oct
5
2
1
28
Sep
25
23
21
18
17
16
15
14
11
10
9
8
4
3
2
1
31
Aug
28
27
26
25
24
21
20
19
18
17
14
13
12
11
10
30
Jul
29
28
27
24
23
22
21
20
17
16
15
14
10
9
8
7
6
3
2
Use a more conservative buffer size when clipping S2-tiles.
Working on reading buffered raster to ensure whole tile extent.
Implemented tiling of a raster into MGRS tiling grid.
Implementing a function to segment a raster into the MGRS grid system.
Implemented a posteriori probability.
Implemented 1d convolution preserving input shape.
Implemented a more efficient method to read hdf4 files.
Added rasterio and h5py, h5netcdf as dependencies.
Implemented option to return model predictions.
Added initial logging to trainer.
Fixed initialization of default dictionaries.
Overwrite number of training mini-batches to target domain.
Iterate over target dataloader during UDA.
Do not plot model training loss/accuracies.
Re-initialize metric tracker with additional uda metrics.
Copy tensors to GPU if available.
Added correct returns to uda forward function.
Fixed typo in function name.
Do not increase weight for uda.
Improved reading routine for hdf files.
Initialize logging with model name.
Use correct state file to save network configuration.
Fixed typo: supervised or unsupervised.
Adjust metric tracker only for unsupervised domain adaptation.
Refactoring transfer learning code.
Fixed model name for supervised transfer learning.
Operational settings.
Added settings for Sparcs target dataset.
Computing cm on GPU requires rewriting sklearn cm. Aborting.
Compute confusion matrix on GPU, if available.
Compute confusion matrix on GPU, if available.
Force Float32 dtype when plotting.
Fixed typo: Labels.
Fixed typo: Missing id.
Copy Tensor to cpu before mapping to target labels.
Just set accuracy as title when plotting samples.
Subdataset does not have sensor property.
Do not store scene inputs and use more descriptive figure title.
Set bigger default font sizes.
Make output path to positional argument.
Loading