Releases: frankkramer-lab/aucmedi
Releases · frankkramer-lab/aucmedi
aucmedi_v0.10.0
🐛 Bug Fixes
6ab16d3- metrics: remove num_classes parameter from F1Score and update requirements.txtbd4cff3- F1Score: correct computation of F1Score in automl training13b4eda- change input_stack from list to tuple to resolve TypeErrorbb93486- replace deprecated layer.output_shape with layer.output.shapeafa71bb- add value=0 to pad function and fix RandomCrop function
♻️ Refactors
dd1b77a- commented out all ViT mentions in files9289d90- tests: commented out ViT-related unittests437d3d0- update imports and comments, remove deprecated dependencies8bbd83c- metrics: delete F1Score in the AutoML training block8e5c66e- NeuralNetwork: remove unused workers parameter23c78e8- NeuralNetwork: remove unused multiprocessing parameter9278360- NeuralNetwork: remove unused batch_queue_size parameterf201b75- F1Score: delete F1Score in AutoML training blocka14af09- replace .hdf5 with .keras formatd591efe- update version in requirements61277cf- rename test_application_multilabel4290352- add border_mode to pad function and cv2 import45eed28- comment out resize parameter to resolve ValueError in testsa6be1c7- add input_shape parameter in NeuralNetwork in each testaccc7c1- add input_shape in NeuralNetwork in each test560f1cb- comment out test_utils_iter test493e5e3- update DataGenerator and architectures for resize=(64, 64, 64)
👷 Build System
81d8db2- updated to version 0.10.0 (commit by @muellerdo)
Contributors
muellerdo
Assets 2
aucmedi_v0.9.0
✨ New Features
7b88203- Architecture: added ConvNeXt 3D architectures (commit by @muellerdo)2cfec6e- Architecture: integrated ConvNeXt 3D architectures into AUCMEDI environment (commit by @muellerdo)cb41966- added multi-modality support (commit by @muellerdo)5fd029c- sitk_loader: allow more dynamic resampling in terms of multi-modality (commit by @muellerdo)979a660- Architectures: removed 2D ResNeXt image architecture support (commit by @muellerdo)76866a9- Architecture: added 2D ResNeXt (50 & 101), again... (commit by @muellerdo)
🐛 Bug Fixes
a7e335c- Ensemble: fixed incorrect callback extension (commit by @muellerdo)78b3078- classes parameter (np.array) (commit by @Sassy-Nazogram)f6945b8- Evaluation: changed datatype from float to object for ROC multi-lists (commit by @muellerdo)6d0df6a- Evaluation: replaced np.object with object due to deprecated (commit by @muellerdo)8cb896f- Evaluation: specified groupby column to be consistent with pandas 2.X (commit by @muellerdo)
♻️ Refactors
4530bb9- AutoML: added Tensorflow Addons F1 to AutoML train again (commit by @muellerdo)e5940db- Tutorials: reverted to TF addons F1 metric in the tutorials (commit by @muellerdo)
✅ Tests
c2ce77b- Architecture: added unittesting for ConvNeXt 3D (commit by @muellerdo)7549dc8- unittesting for 3D multi-modality data loading (commit by @muellerdo)4414024- sitk_loader: improved unittesting for resampling of multi-modality nii/mha volumes (commit by @muellerdo)
👷 Build System
800713b- increased package version 3D-classification-models to 1.0.7 (commit by @muellerdo)4caf08d- Updated to version 0.8.2 (commit by @muellerdo)7750c96- Tensorflow: incrased package version to 2.11.1 (commit by @muellerdo)8837e06- Tensorflow-Addons: removed TF-addon and keras-applications support (commit by @muellerdo)2ce29b2- updated dependency list (commit by @muellerdo)8b7da52- Python: dropped Python3.8 for CI and added 3.9 & 3.10 (commit by @muellerdo)6cbe6e9- Python: removed Python 3.12 again from CI as there is no TF1.4 support there (commit by @muellerdo)2de4ba3- Tensorflow-Addons: added TFaddons back due to requirements for Keras ViT.. (commit by @muellerdo)fc11a51- Keras-Applications: added also back Keras Applications (and 2D ResNeXt) as classification-models-3D require it... (commit by @muellerdo)6990dd5- Python: converted python version to strings to be 3.10 instead of 3.1 (commit by @muellerdo)7152f87- plotnine: upgraded plotnine to version 0.12.0 (commit by @muellerdo)5db39c1- plotnine: fixed version to 0.12.4 (commit by @muellerdo)c48c4c9- scikit: updated scikit-learn & scikit-image (commit by @muellerdo)ba40783- Pillow: updated Pillow to 10.2.0 for security (commit by @muellerdo)da8c660- updated to version 0.9.0 (commit by @muellerdo)
📝 Documentation Changes
812bb88- Tutorials: fixed incorrect link - solves #200 (commit by @muellerdo)
🔧 Chores
444c11d- updated copyright to 2023 (commit by @muellerdo)729a275- merged dev branch to ConvNeXt3D PR (commit by @muellerdo)22d5415- updated copyright to 2024 (commit by @muellerdo)a155af8- updated copyright on website (commit by @muellerdo)
Contributors
muellerdo and 3410-Sassy
Assets 2
aucmedi_v0.8.1
♻️ Refactors
80566ed- DataGenerator: removed overlooked debugging sys.out print (commit by @muellerdo)
👷 Build System
ecef046- Updated to version 0.8.1 (commit by @muellerdo)
Contributors
muellerdo
Assets 2
aucmedi_v0.8.0
✨ New Features
2e8f2a9- DataGenerator: rework deprecated keras.Iterator to keras.Sequence - Closes #76,#189 (commit by @muellerdo)
🐛 Bug Fixes
8f1a2e8- Evaluation: updated deprecated np.int to np.int8 for boolean to integer matrix conversion (commit by @muellerdo)
✅ Tests
c70fc4c- NeuralNetwork: testing for repeatition capabilities in datagenerators (commit by @muellerdo)5568cfa- DataGenerator: added iteration & repetition testing as well as batch access (commit by @muellerdo)
👷 Build System
c27c7f1- Codecov: added config file with threshold of 1% (commit by @muellerdo)dc264bf- Updated to version 0.8.0 (commit by @muellerdo)
📝 Documentation Changes
973ff46- README: fixed package build status badge (commit by @muellerdo)
Contributors
muellerdo
Assets 2
aucmedi_v0.7.3
💥 BREAKING CHANGES
-
due to
b38dd6d- updated various package versions (commit by @muellerdo):updated various package versions
✨ New Features
82c3e3d- Architecture: added ConvNeXt - closes #101 (commit by @muellerdo)5d94fbe- Architecture: added ConvNeXt to architecture library (commit by @muellerdo)
✅ Tests
b379dc2- Architecture: added unittesting for ConvNeXt (commit by @muellerdo)
👷 Build System
34dc3ca- TensorFlow: increased tensorflow to at least v2.10 (commit by @muellerdo)b38dd6d- updated various package versions (commit by @muellerdo)d089d88- TensorFlow-Addons: fixed typo in version (commit by @muellerdo)efb6297- Updated to version 0.7.3 (commit by @muellerdo)
📝 Documentation Changes
1e96578- Example: removed template pipeline of standardized framework example (commit by @muellerdo)6985b72- Example: added link to Medium blog article (commit by @muellerdo)a3781fa- Example: added manuscripts to framework examples (commit by @muellerdo)e867b34- Reference: added ref to application preprint paper (commit by @muellerdo)
Contributors
muellerdo
Assets 2
aucmedi_v0.7.2
👷 Build System
dbbdbd6- PRs: added code quality CI to PRs targeting master (commit by @muellerdo)0da9e1b- protobuf: increased to v3.20.2 as security update (commit by @muellerdo)41cc227- matplotlib: maxed to v3.5.3 due to plotnine dependency (commit by @muellerdo)73304f7- Updated to version 0.7.2 (commit by @muellerdo)
📝 Documentation Changes
5405268- Website: fixed a typo in the status bar (roadmap) (commit by @muellerdo)f3c3186- Classifier: enhanced the example in the API ref (commit by @muellerdo)1b0a51f- add links and fix example (commit by @SherlockMones)90e4651- fix links (commit by @SherlockMones)f555fce- stacking: fix example (commit by @SherlockMones)0250281- Example: added multiple API examples and tutorials (commit by @muellerdo)
aucmedi_v0.7.1
✨ New Features
f02195e- Evaluation: added show option for dataset evaluation plots (commit by @muellerdo)
🐛 Bug Fixes
976369a- Evaluation: fix incorrect AUC computation - closes #165 (commit by @muellerdo)cbe64f9- Evaluation: add non-complete data support for confusion matrix - closes #167 (commit by @muellerdo)81173d3- XAI: replace separator chars with . for xai_decoder result saving - fixes #169 (commit by @muellerdo)
✅ Tests
ed10587- Evaluation: added edge case support of non-complete data for confusion matrix (commit by @muellerdo)c6dae5d- XAI: added test for xai_decoder with directory interface (commit by @muellerdo)
👷 Build System
abb5601- updated to version 0.7.1 (commit by @muellerdo)
📝 Documentation Changes
168c1e2- Various updates in README and addition of tutorial placeholder on website (commit by @muellerdo)96bca31- README: removed getting started chapter title artefact (commit by @muellerdo)5d336f4- Subfunctions: removed artefact init docs of ABC (commit by @muellerdo)173b529- XAI: added return docs for xai_decoder - closes #143 (commit by @muellerdo)96765ed- Contributing: enhanced contributing guideline (commit by @muellerdo)
aucmedi_v0.7.0
✨ New Features
21323b5- Ensemble: Implemented Composite (cross-validation based stacking) (commit by @muellerdo)20f61f4- AutoML: Implemented function block - training (commit by @muellerdo)099e105- AutoML: Added multilabel support for function block training (commit by @muellerdo)74fcfbd- AutoML: added metadata storing in training block (commit by @muellerdo)5fab10a- AutoML: Implemented function block prediction (commit by @muellerdo)2a2e9e9- AutoML: Implemented XAI option in code block inference (commit by @muellerdo)5b4c095- Evaluation: added support for transfer learning in fitting evaluation (commit by @muellerdo)47aff85- AutoML: added evaluation plot generation in block function training (commit by @muellerdo)fad6f2d- AutoML: implemented block function evaluation (commit by @muellerdo)f7ccf82- AutoML: Added CLI parser for Training (commit by @muellerdo)c521785- AutoML: added CLI interface for YAML (commit by @muellerdo)0b52a90- AutoML: added CLI interface for prediction (commit by @muellerdo)6a54157- AutoML: added CLI interface for evaluation (commit by @muellerdo)2f90edd- AutoML: added more sanity checks to evaluation function block (commit by @muellerdo)7f66f48- performance: add display option for charts (commit by @SherlockMones)87ae17a- io_interfaces: add file extension to index_list (commit by @SherlockMones)2d563b0- AutoML: added binary packaging via entry script (commit by @muellerdo)2c5ac70- data_processing: Improve documentation (commit by @SherlockMones)3566791- AutoML: reworked AutoML CLI arguments - closes #151 (commit by @muellerdo)96c8d44- AutoML: dockerized AUCMEDI AutoML - closes #150 (commit by @muellerdo)c53819f- evaluation: display fitting progress (commit by @SherlockMones)4711894- Evaluation: added full transfer learning support for fitting evaluation - closes #157 (commit by @muellerdo)cbc2510- Stacking: added automated standardize and resize parameter estimation for DataGenerator (commit by @muellerdo)
🐛 Bug Fixes
ea04aad- Ensemble: Added missing library import in Stacking & Bagging (commit by @muellerdo)cc53a57- AutoML: changed incorrect grayscale parameter of DataGenerator in training function block (commit by @muellerdo)8b973d2- AutoML: some hotfixes for AutoML XAI support (commit by @muellerdo)7c78357- AutoML: incorrect output path as directory creation (commit by @muellerdo)4a23572- AutoML: added more config parsing for CLI (3D shape, input and output path) (commit by @muellerdo)c31355d- Evaluation: catch no-possible AUROC computation (commit by @muellerdo)f4b165c- AutoML: fixed multiple architecture passing for advanced training (commit by @muellerdo)6729798- Evaluation: fixed unequal length of training epochs by multi model pipelines (commit by @muellerdo)c5f9712- AutoML: forget to exluce yaml parser function in main runner (commit by @muellerdo)05169cf- XAI: added missing inverse of occlusion sensitivity heatmap (commit by @muellerdo)87e3f9d- evaluation: fix evaluate training charts (commit by @SherlockMones)e474282- Evaluation: added support for only transfer learning epochs in evaluate_fitting() (commit by @muellerdo)
⚡ Performance Improvements
c79110f- Ensemble: Replaced python built-in multiprocessing to pathos multiprocessing (commit by @muellerdo)816fea3- Ensemble: Re-initialize NeuralNetworks in processes to allow custom loss functions (commit by @muellerdo)ab6c0bc- Ensemble: allow dumping of ensemble output into already existing directories (commit by @muellerdo)d3988a5- AutoML: Various improvements of the function block training (commit by @muellerdo)e5b8ae2- AutoML: added sanity checks for multi_label supported metalearners (commit by @muellerdo)2b9b355- Evaluation: call explicit categorical parsing of class names for omnious class names (numbers) (commit by @muellerdo)834799e- AutoML: switched csvlogger to append mode (commit by @muellerdo)850a6a1- AutoML: adjusted t...
aucmedi_v0.6.1
👷 Build System
bc2104f- Unittesting: Renamed GitHub Action workflow for package building (commit by @muellerdo)bc34d09- Code-Quality: Refactored code coverage GitHub Action workflow (commit by @muellerdo)34bde90- Code-Quality: Added GitHub Action commitlint for commit convention checking (commit by @muellerdo)f226d89- Code-Quality: currently disabling commitlint (commit by @muellerdo)43f6c8a- Code-Quality: added configuration file for commitlint (commit by @muellerdo)67c6d9b- Updated to version 0.6.1 for Zenodo release (commit by @muellerdo)
aucmedi_v0.6.0
✨ New Features
de02965- Bagging: Implemented training process for Bagging (also closes #121) (commit by @muellerdo)c7cbc8e- Bagging: Implemented Bagging training in separate Processes to avoid any garbage collection issues (commit by @muellerdo)c9ed177- Bagging: Implemented Bagging inference functionality (commit by @muellerdo)f78bba3- Bagging: Implemented Bagging dump (result storage) functionality (commit by @muellerdo)41c4114- Bagging: Implemented Bagging load (re-using result directory) functionality (commit by @muellerdo)7014130- Metalearner: Started working on Metalearners for later usage in Stacking (commit by @muellerdo)cdf9f54- Metalearner: Implemented abstract base class for Metalearner (commit by @muellerdo)168116f- Stacking: Implemented ensemble learning technique Stacking (commit by @muellerdo)633af82- Metalearner: Implemented Metalearner Logistic Regression (commit by @muellerdo)726c6f2- Metalearner: Implemented Metalearner: Naive Bayes (commit by @muellerdo)b7d5548- Metalearner: Added Metalearner Support Vector Machine (commit by @muellerdo)d07b12d- Metalearner: Added Metalearner Gaussian Process (commit by @muellerdo)85b861b- Metalearner: Added Metalearner Decision Tree (commit by @muellerdo)54d6f4e- Metalearner: Added Metalearner Best Model (commit by @muellerdo)162b977- Metalearner: Added Metalearner Weighted Mean (commit by @muellerdo)be429f6- Metalearner: Added Random Forest - Contributes to #129 (commit by @muellerdo)817e02d- Metalearner: Added k-Nearest Neighbors - Contributes to #129 (commit by @muellerdo)b66de88- Metalearner: Added MLP Neural Network - Contributes to #129 (commit by @muellerdo)205dd12- Aggregate: Added Global Argmax - Contributes to #129 (commit by @muellerdo)09d4b13- Classifier: Removed two_dim parameter from Classifier (commit by @muellerdo)2c8f13a- Architectures: Started working on Vision Transformer (commit by @muellerdo)33068ac- Architectures: Implemented Vision Transformer Architecture (commit by @muellerdo)32a43d1- Utils: Added utility function for parsing CSVLogger fitting file to fitting dictionary (commit by @muellerdo)da69ac2- Evaluation: Implemented fitting curve plotting (commit by @muellerdo)393274f- Evaluation: Started implementing performance evaluation (commit by @muellerdo)2962aa3- Evaluation: Fully implemented performance evaluation (commit by @muellerdo)1c80fdd- Evaluation: Added unittesting for performance evaluation (commit by @muellerdo)0cdc6bd- Evaluation: Implemented comparison evaluation function (commit by @muellerdo)4f90c5c- Ensemble: Added return_ensemble feature for Stacking and Bagging predictions. (commit by @muellerdo)c2c1a33- Evaluation: Added returning merged and gain dataframe in Comparison Evaluation (commit by @muellerdo)8da7c53- Evaluation: Added dataset analysis evaluation function (commit by @muellerdo)
🐛 Bug Fixes
d764ff9- Bagging: Force setting multiprocessing start option which solves multi-assignment (commit by @muellerdo)00cefe4- Bagging: Removed unittests for deprecated model_list variable in Bagging initialization (commit by @muellerdo)275b0b8- Bagging: Fixed incorrect DataGenerator parameter mapping in predict function (commit by @muellerdo)72794df- Metalearner: Fixed postprocessing of Decision Tree prediction (commit by @muellerdo)546c07b- Website: Fixed 'funky' first item on landing page - Closes #106 (commit by @muellerdo)2584e78- Website: Hotfix of new introduced bug with previous commit (commit by @muellerdo)e44916b- Architectures: Fixed incorrect input shapes for B16 and B32 Vision Transformer Architecture (commit by @muellerdo)1b554cc- Evaluation: Fixed suffix=None usage in evaluate_fitting() (commit by @muellerdo)c8ce20d- Classifier: Removed deprecated two_dim parameter from unittesting (commit by @muellerdo)