75d07b96e8
ENH: Autoupdate hooks [pre-commit.ci] ( #737 )
...
* [pre-commit.ci] pre-commit autoupdate
updates:
- [github.com/pre-commit/pre-commit-hooks: v4.0.1 → v4.3.0](https://github.com/pre-commit/pre-commit-hooks/compare/v4.0.1...v4.3.0 )
- [github.com/PyCQA/flake8: 3.9.2 → 4.0.1](https://github.com/PyCQA/flake8/compare/3.9.2...4.0.1 )
- [github.com/pre-commit/mirrors-autopep8: v1.5.7 → v1.6.0](https://github.com/pre-commit/mirrors-autopep8/compare/v1.5.7...v1.6.0 )
* Update .pre-commit-config.yaml
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Fernando Pérez-García <fperezgarcia@microsoft.com>
2022-07-07 16:03:39 +01:00
cf8bd5e906
BUG: Ensemble HeadAndNeck model encounters 'CUDA out of memory' ( #754 )
...
Fix cuda out of memory for ensembles
2022-06-30 17:00:59 +01:00
38f830374e
ENH: Move hi-ml and fastMRI to PyPI packages ( #752 )
...
* 🔥 💥 Remove submodules
* ➕ Add packages to primary dependencies
* 📌 ⬆️ Upgrade python to 3.8 and lock
* ✅ Fix fastMRI test
* 💚 ✅ Fix tests and CI workflows
* 🏷️ Fix MyPy and Flake8 issues
* ⬆️ Update windows dependencies
* 📝 Add fastMRI reference
* ⚰️ Remove references to old submodules
2022-06-23 09:44:54 +01:00
69a5e8cd2c
DOC: Update environment documentation ( #748 )
...
* 📝 Update environment documentation
* 📝 Add quick setup, add review comment changes
2022-06-22 10:57:30 +00:00
b42043f31a
ENH: Remove cucim + change lock script line endings ( #750 )
...
📌 Remove cucim + change lock script line endings
2022-06-21 22:45:23 +01:00
106c101ff7
DOC: Improve doc for checkpointing ( #747 )
2022-06-21 17:09:24 +01:00
6a103f14ea
ENH: Sample all classes with equal probability in hippocampus config ( #746 )
2022-06-20 15:40:33 +01:00
cf1de78ae5
ENH: Add hippocampus segmentation config file ( #739 )
...
* Adding hippocampus segmentation config file
2022-06-14 11:50:06 +01:00
a9fd52c2c9
BUG: Fix missing channels dimension in normalization ( #701 )
...
* Fix missing channels dimension in normalization
* Update CHANGELOG
* Add test for 3D and 4D input images
* Move conversion to NumPy array
2022-06-13 13:47:27 +01:00
edc72eda8c
STYLE: Fix pre-commit errors ( #736 )
...
* STYLE: Apply some autopep8 fixes
* Fix more pre-commit errors
2022-06-07 09:14:00 +01:00
ad9aac5413
DOC: Improve documentation for dataset creation ( #723 )
2022-06-06 16:12:10 +00:00
796594f6f7
STYLE: Disable pre-commit CI autofixes
2022-06-01 21:46:49 +01:00
92d94799f2
ENH: Add Environment Locking ( #735 )
...
* ✨ Add environment locking script
* 📝 ✨ Finish script, add documentation
* 🐛 Change Windows env file in workflow
* 📝 🐛 Add review changes + fixes
* 🚧 Temporarily alter tests and conda channels
* 🚧 Add logging
* ✅ Fix TestSubmodule env file
* 🔥 Delete env test
* 🧑💻 Add warning to environment.yml
* 📝 ⚰️ Update based on review comments
* 📝 Add final changes
2022-06-01 10:05:54 +00:00
c7e4a3e945
Auto add issues to GitHub Project ( #721 )
...
Adds a new workflow that automatically adds any new issues to the InnerEye-OSS GitHub Project
2022-05-26 09:19:48 +01:00
a3a066e4fe
ENH: Add a structured template for issues ( #727 )
2022-05-24 11:47:44 +01:00
bf075678ca
BUG: Fix broken build pipeline by updating package versions ( #725 )
...
* ⬇️ Fix cloudpickle to appropriate version range
* ⬆️ Upgrade InnerEye-DICOM-RT Dependency
2022-05-18 09:13:21 +01:00
a15a1a2ce3
DOC: Add part of the API to the Sphinx docs ( #718 )
...
* Add part of the API to the Sphinx docs
* Improve layout of landing page
* Remove redundant caption
* Include all documentation in Sphinx and fix docstrings
* Remove redundant options
2022-04-19 15:04:17 +01:00
c5b16e5d62
ENH: Customized VSCode Window Color ( #712 )
2022-04-13 14:47:05 +01:00
14b968c282
DOC: Add a smoke-test model for AzureML setup ( #708 )
...
- Details how to test AzureML without uploading data.
- Also sneaked in an update of hi-ml to fffbb0f8
2022-04-13 14:05:09 +01:00
6aab3b1f93
STYLE: Ensure line breaks at the end of all files ( #717 )
...
* Add line breaks at the end of all files
* Handle last line in VS Code settings
2022-04-04 15:08:43 +01:00
2edbdc8d64
STYLE: Trim trailing whitespaces with pre-commit ( #715 )
2022-04-04 10:27:38 +01:00
760cbe0305
ENH: Enforce prefixes for pull request titles ( #713 )
...
Enforce prefixes for pull request titles
Co-authored-by: Anton Schwaighofer <antonsc@microsoft.com>
2022-03-30 15:18:28 +01:00
5aef988b96
Ignore Sphinx build folders ( #698 )
2022-03-30 14:47:28 +01:00
3eaf2f9053
Deprecate manual changelog ( #707 )
2022-03-29 13:43:50 +01:00
f1ae24aa69
Update hi-ml submodule ( #709 )
2022-03-29 08:56:46 +01:00
b2ae826653
Import submodules to sys.path to fix autodoc's warning ( #704 )
2022-03-23 21:32:00 +00:00
6791dce4c3
Automatically and linearly scale the learning rate of the SSL encoder to the number of GPUS ( #667 )
...
The learning rate is now linearly scaled by the number of GPUs available, e.g., lr = 0.001, 8 GPUs are available => lr=0.008.
I tested it for SimCLR and BYOL (https://ml.azure.com/experiments/id/81fa8775-1a25-47ae-9fe7-c13a6b91a421?wsid=/subscriptions/db9fc1d1-b44e-45a8-902d-8c766c255568/resourceGroups/innereyerg/providers/Microsoft.MachineLearningServices/workspaces/innereye4ws&tid=72f988bf-86f1-41af-91ab-2d7cd011db47 ).
In both cases the results are as expected:
SimCLR: More GPUs give better representations due to larger batches == more negative eaxmples.
BYOL: More GPUs speed up the training but do not give better results.
I did not add any test since we have these two tests already: test_simclr_num_gpus() and test_simclr_num_nodes()
2022-03-22 10:23:39 +00:00
45e7d5ff4d
Fix some Sphinx warnings ( #699 )
...
* Fix shell lexer name
* Update CHANGELOG
* Fix CHANGELOG
* Fix "html_static_path entry '_static' does not exist"
* Clean up preprocess script
* Fix link to InnerEye-DataQuality
* Use shutil.copy to copy files
* Remove extra info from CHANGELOG
* Fix broken link to LICENSE
* Fix lexer name for YAML
* Remove colons from headers
* Fix InnerEye module not being found
2022-03-22 09:46:15 +00:00
95d8b72ae6
Specify `formatter_class=argparse.ArgumentDefaultsHelpFormatter` in `generic_parsing.py` to show default values for arguments in help message ( #689 )
2022-03-15 17:27:01 +00:00
f0d2337d1b
Improve setup for running the HelloWorld model in AzureML ( #693 )
2022-03-15 14:07:28 +00:00
63fb8688b0
Update TorchIO version ( #677 )
...
* Update TorchIO version
Resolves #672 .
2022-03-14 16:50:29 +00:00
d929ccbd46
Ensure the shape of input patches is compatible with model constraints ( #682 )
...
* Ensure patches shape is compatible with model constraints
* Remove test for images that are too small
2022-03-10 17:01:25 +00:00
3c919dca14
Replace data quality folder with a link to the commit ( #692 )
...
* Move to commit link
* Remove build
* Update paper link
* Update README
* Add changelog
2022-03-10 09:14:25 +00:00
1606729c7a
Clean up legacy code ( #671 )
...
* Remove rnns
* Fix flake8
* Edit README
* Edit README
* Remove sequence
* Remove sequence
* Fix all
* Remove more
* Remove ignore
* Fix tests
* Undo config
* Fix config
* Revert pycharm
* Fix tests
* Undo outputlogger
* Fix flake8
* Fix ignore file
* Revert hi-ml
* Disable fail on alert
2022-03-09 14:53:12 +00:00
8a78ec8c1e
Pad model outputs if they are smaller than the inputs ( #681 )
2022-03-09 13:32:26 +00:00
5336a67f66
Fix Markdown linting errors ( #684 )
2022-03-08 16:48:53 +00:00
b1f3ef54d6
Fix missing separator error in docs Makefile ( #683 )
2022-03-07 18:00:34 +00:00
6b62d8cbcd
Print out name of log level, not only the number ( #678 )
2022-03-07 10:11:47 +00:00
501f3920c6
Add FP and TN to DeepMIL outputs and configure outputs for multi-class classification ( #679 )
2022-03-04 16:16:49 +00:00
a4c82df78e
Updated README.md ( #680 )
...
Add acknowledgements section
Co-authored-by: Fernando Pérez-García <fperezgarcia@microsoft.com>
2022-03-03 16:46:05 +00:00
e984554c9e
Fix DeepMIL metrics bug ( #674 )
...
* Fix DeepMIL metrics input bug
* Add first version of metrics tests
* Update submodule
* Add test for DeepMIL metrics inputs
* Clean-up and update submodule
* Update changelog
* Upgrade mlflow due to Component Governance warning
2022-03-01 10:10:13 +01:00
d7e5d8b5e5
Replace RadIO with TorchIO for patch-based inference ( #666 )
...
* Replace RadIO with TorchIO
* Ensure patches are float32 for forward pass
* Update changelog
* Ignore some types to fix mypy errors
* Remove APEX from conda environment in docs example
Co-authored-by: Javier <jaalvare@microsoft.com>
2022-02-23 10:28:11 +00:00
e2ec5cc839
Add subsampling transform and mean pooling ( #656 )
...
* Add subsampling transform
* Add option to allow_missing_keys for Subsampled
* Add dropout param to BaseMIL
* Add docstring and tests for Subsampled
* Update changelog
* Update to hi-ml with mean pooling
* Enable mean pooling in DeepMIL
* Add/refactor mean pooling tests
* Update changelog
* Update to latest hi-ml with mean pooling
2022-02-21 11:24:14 +00:00
1600ef3ddf
Fix DeepMIL for TCGA CRCK dataset ( #659 )
...
While we updated DeepMIL for the Panda dataset to work with the latest changes, we did not update DeepMIL for the TCGA CRCK dataset.
This PR updates how the caching of the encoded tiles is done and how the checkpoints of the DeepMIL model is saved and loaded.
No additional tests are required since these are the same functions that we use for the Panda dataset. For all of them a test already exists.
Last, the PR updates the cudatoolkit version, Anton and I found that this is the root cause for all our problems with ddp
2022-02-16 09:50:59 +00:00
914a89383d
Enable fine-tuning in Deepmil ( #650 )
2022-02-09 10:01:57 +00:00
eda76357f0
Add dropout to DeepMIL and fix feature extractor setup ( #653 )
...
* Add dropout to DeepMILModule, with param in BaseMIL
* Fix feature extractor setup for torchvision models
2022-02-07 13:09:04 +00:00
d617c8107c
Re-enable Linux build ( #655 )
...
Also refactored tests that caused test discovery to slow down
2022-02-04 15:41:42 +00:00
a6b15166b7
Fix for stuck Linux build: Move pytest to Windows ( #652 )
...
Also renamed many build legs so that they can be found more easily in the UI.
2022-02-04 11:20:03 +00:00
b8fe0eb2f1
Fix in _convert_to_tensor_if_necessary so that also np.array get converted correctly ( #649 )
2022-02-01 14:53:41 +00:00
4fdeed23e4
Workaround for bug in PL 1.5.5: CombinedLoader cannot be used with DDP for training data ( #646 )
...
Fix for https://github.com/PyTorchLightning/pytorch-lightning/issues/11632
2022-02-01 14:25:21 +00:00
pre-commit-ci[bot]
Melissa Bristow
Peter Hessey
Anton Schwaighofer
Anna Schroder
Fernando Pérez-García
Jacopo Teneggi
maxilse
Javier
Harshita Sharma
Dr Kenji Takeda
Daniel Coelho de Castro
Fernando Pérez-García