pFL combinations, new checkpointing modules, early stopping, more nnUnet integration, way better smoke tests

What's Changed

• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in #267
• Modify LR Scheduler + Test by @jewelltaylor in #268
• try except block for dataset.json in case nnunet_raw is unavailable by @scarere in #266
• Bump waitress from 3.0.0 to 3.0.1 by @dependabot in #269
• Address Shawns CR on previous PR by @jewelltaylor in #270
• Elapsed time was backwards by @scarere in #271
• modified progress bar formatting to inherit from flwr formatting by @scarere in #273
• Modify curr step of LR Scheduler to start at 0 instead of 1 by @jewelltaylor in #274
• Fix typing on assertion for ReportsManager by @jewelltaylor in #276
• Changes to Support Expanded Experimentation with FedDG-GA by @emersodb in #252
• Bump pypa/gh-action-pypi-publish from 1.10.3 to 1.11.0 by @dependabot in #277
• Adaptive and Fed DG-GA PFL Experimentation by @emersodb in #251
• Add More Logging To FedDG-GA Strategy by @emersodb in #272
• Updating the FLamby installation instructions to reflect some changes on their end. by @emersodb in #275
• Lowercasing torch device variables throughout library when not global by @emersodb in #279
• Pinning a source of randomness in server-side aggregation. by @emersodb in #278
• First pass at optionally shutting down server (and remaining clients) on failures by @emersodb in #280
• Adding a touch of logging for checkpointing errors by @emersodb in #281
• Bump pypa/gh-action-pypi-publish from 1.11.0 to 1.12.2 by @dependabot in #284
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in #282
• More Masked Layers for FedPM by @yc7z in #283
• Fixing infinite base server test by @emersodb in #289
• Moving relevant files to use the future annotations import for 'in-class' return values by @emersodb in #286
• Modified some of the reporting keys and overhauled the wandb reporter by @scarere in #288
• Small update to FLamby install instructions to avoid cluster failures. by @emersodb in #287
• Bump aiohttp from 3.10.3 to 3.10.11 by @dependabot in #290
• Name change for server folder to servers by @emersodb in #292
• Add resume wandb reporter by @jewelltaylor in #293
• Update deep mmd client and loss implementation by @sanaAyrml in #285
• Custom nnunet trainer by @jewelltaylor in #296
• Bump mlflow from 2.15.1 to 2.16.0 by @dependabot in #297
• Bump pypa/gh-action-pypi-publish from 1.12.2 to 1.12.3 by @dependabot in #302
• Converting to masked layers bug fix by @yc7z in #303
• Server storing FL configs and Consolidating Base Server Functionality by @emersodb in #294
• Add max_num_validation_steps member of config and client and related … by @jewelltaylor in #304
• Nnunet amp integration by @jewelltaylor in #305
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in #306
• Upgrading peotry lock to include jinja update. by @emersodb in #308
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in #310
• Consolidating model and state checkpointing on the client and server sides. by @emersodb in #298
• Mypy Typing Migration and Some typo fixes by @emersodb in #299
• A few fixes to our Examples by @emersodb in #300
• Keras vulnerability ignore until there is a fix by @emersodb in #312
• [CI] Add github action for creating Github release by @nerdai in #311
• Small download script changes for RxRx1 by @emersodb in #315
• Smoke test bug fix for Andrei with the state checkpointers by @emersodb in #316
• [CU-868byqhk5] Fix smoke tests for GPU by @nerdai in #314
• Pin PyArrow to v^17.0 by @nerdai in #317
• Rxrx1 research running scripts by @sanaAyrml in #309
• Move assert statements out of run_smoke_test and into the actual test (for graceful shutdown in case of failure) by @nerdai in #318
• Add early stop module by @sanaAyrml in #301
• Bump pypa/gh-action-pypi-publish from 1.12.3 to 1.12.4 by @dependabot in #324

New Contributors

• @nerdai made their first contribution in #311

Full Changelog: v0.2.2...v0.3.0

Second patch fix for a few small straggler errors

What's Changed

• Move reporting start_time in servers to before fit call by @jewelltaylor in #261
• Nnunet server patch by @jewelltaylor in #262
• Make necessary changes to model merge server and client with new repo… by @jewelltaylor in #263
• Turning follow imports on in mypy for better coverage by @emersodb in #264

Full Changelog: v0.2.1...v0.2.2

Small patch release to include the proper core dependencies in the library toml

What's Changed

• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in #259
• Update dependencies by @jewelltaylor in #260
• PR4: Add deep_mmd_loss files by @sanaAyrml in #170

Full Changelog: v0.2.0...v0.2.1

New nnunet integrations, better heterogeneous sampling, MMD constraints, reporter refactors

What's Changed

• Proposed changes to the mkmmd loss function by @emersodb in #102
• Update MK-MMD Loss Suggestions and Tests by @emersodb in #181
• CIFAR Experiments PR Suggestions by @emersodb in #184
• Changed nnunet LR Schedule and fixed memory leak by @scarere in #222
• Bump nltk from 3.8.1 to 3.9 by @dependabot in #223
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in #224
• Created reset signal handler decorator by @scarere in #225
• Adding a README to the examples folder. by @emersodb in #226
• Bump pypa/gh-action-pypi-publish from 1.9.0 to 1.10.0 by @dependabot in #229
• Update requirements by @jewelltaylor in #230
• Update scaffold by @jewelltaylor in #231
• Nnunet core refactor by @scarere in #227
• Bump pypa/gh-action-pypi-publish from 1.10.0 to 1.10.1 by @dependabot in #233
• Basic client checkpointing and LR Scheduling by @jewelltaylor in #234
• Add updates for smaller CRs by David by @jewelltaylor in #236
• Adding in Dirichlet allocation partitioner by @emersodb in #237
• Expanding ClientCheckpointModule to accept checkpointer sequences by @emersodb in #239
• Quick update for partitioner to also get a prior rather than sampling by @sanaAyrml in #242
• PR2: Update mkmmd loss by @sanaAyrml in #169
• Bump pypa/gh-action-pypi-publish from 1.10.1 to 1.10.2 by @dependabot in #244
• PR1: Add mkmmd loss by @sanaAyrml in #168
• Nnunet server checkpointing by @jewelltaylor in #241
• Revert "Nnunet server checkpointing" by @jewelltaylor in #245
• Nnunet server checkpointing by @jewelltaylor in #246
• Adding in synthetic dataset generators based on the original FedProx paper by @emersodb in #238
• Cleanup picai folder by @jewelltaylor in #247
• Fix launch scripts by @jewelltaylor in #248
• PR3: Add cifar10 experiments by @sanaAyrml in #171
• Adaptive Ditto and MR-MTL by @emersodb in #235
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in #249
• Bump pypa/gh-action-pypi-publish from 1.10.2 to 1.10.3 by @dependabot in #250
• Adding fit end and evaluate end times in the client metrics by @lotif in #253
• Fixing the tests for Mk-MMD loss such that they still pass properly on a GPU and CPU by @emersodb in #255
• Reporting restructure by @scarere in #254

Full Changelog: v0.1.17...v0.2.0

New methods, strategies, library support, and more.

There are several big additions in this release.

• We now support federated model merging without requiring training
• The server-side strategy FLASH is now supported
• Native support for FedPM has been implemented
• FENDA+Ditto is now available
• FedSimCLR is now supported for self-supervised contrastive learning.
• A ton of support for nnUnet is now in the library.

What's Changed

• Fix some bugs that slipped through in CR and trying to make type chec… by @jewelltaylor in #172
• Bump urllib3 from 2.2.1 to 2.2.2 by @dependabot in #176
• Bump scikit-learn from 1.4.1.post1 to 1.5.0 by @dependabot in #178
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in #175
• Fixing Some Pip Audit Issues by @emersodb in #180
• Fix picai channel ordering by @jewelltaylor in #179
• Bump pypa/gh-action-pypi-publish from 1.8.14 to 1.9.0 by @dependabot in #177
• FLASH implementation by @Justinyangjy in #164
• Picai transfer learning by @scarere in #174
• simple scripts for eval and pred with nnunet by @scarere in #183
• FedSimCLR by @jewelltaylor in #173
• add setup_client call in get_parameters if not initialized by @jewelltaylor in #186
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in #187
• Picai monai by @scarere in #188
• Bump certifi from 2024.6.2 to 2024.7.4 by @dependabot in #190
• Update flwr 1.7 to 1.9 by @jewelltaylor in #189
• Fenda+Ditto model by @Justinyangjy in #182
• added epoch hook by @scarere in #193
• Skin cancer dataset by @Justinyangjy in #185
• Bump sentry-sdk from 2.5.1 to 2.8.0 by @dependabot in #194
• Fl nnunet by @scarere in #192
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in #196
• Added epoch hook for lr scheduling and made client logs look better by @scarere in #195
• Model Merge Functionality by @jewelltaylor in #191
• FedPM by @yc7z in #197
• Ft/gemini scripts by @fatemetkl in #206
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in #203
• Nnunet test by @scarere in #199
• Nnunet ensemble by @scarere in #198
• Modified base server to accept sequences of checkpointers by @scarere in #205
• Bump pypa/gh-action-pip-audit from 1.0.8 to 1.1.0 by @dependabot in #219
• Updating the README and bump Version Number by @emersodb in #220
• Added progress bar and fixed bugs by @scarere in #204
• decreased nltk version from ^3.8.1 to ^3.8.0 by @scarere in #221

New Contributors

• @scarere made their first contribution in #174

Full Changelog: v0.1.16...v0.1.17

Client Simplifications, PICAI preprocessing, README updates.

What's Changed

• Address mypy issues for DPScaffoldClient by @Justinyangjy in #147
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in #148
• Small update to the readme to include new capabilities added by @emersodb in #150
• Bump requests from 2.31.0 to 2.32.0 by @dependabot in #153
• Ignoring W and B vulnerability for now by @emersodb in #154
• part 1 changes for adding training set capabilities for client side evaluation by @Justinyangjy in #151
• Adding Logo Stuff! by @emersodb in #158
• nnUNet PICAI by @jewelltaylor in #155
• Part 2 changes for test dataset by @Justinyangjy in #156
• Part 1: Extracting a small code refactor from Sana's other work by @emersodb in #152
• Part 2: Creating New Loss Function Classes by @emersodb in #159
• Part 3: Migrating Moon Clients to use new losses by @emersodb in #160
• Part 4: Addition of a loss config and FLamby client simplification by @emersodb in #161
• Part 5: Officially splitting FENDA and PerFCL clients by @emersodb in #162
• Add tests for picai transforms and other test fixes by @jewelltaylor in #165
• Part 6: Adding the Constrained FENDA client by @emersodb in #163
• Bump tornado from 6.4 to 6.4.1 by @dependabot in #166

New Contributors

• @Justinyangjy made their first contribution in #147
• @pre-commit-ci made their first contribution in #148

Full Changelog: v0.1.15...v0.1.16

FedRep, MR-MTL, FedDG-GA, upgrading torch, flwr versions and more!

What's Changed

• Ditto FLamby implementation scripts. by @emersodb in #112
• Small fix for when sampling dataloaders produce an empty batch. by @emersodb in #114
• Fixing two small typos by @emersodb in #118
• Fixing an issue with Opacus modified models and optimizers. by @emersodb in #119
• Upgrading pycyclops to ^0.2.5 and removing exceptions on pip-audit by @lotif in #121
• Update torch, flwr and other requirements by @jewelltaylor in #120
• Displaying all the smoke tests metrics errors at once by @lotif in #122
• Updating the README to include Ditto by @emersodb in #123
• AG News Experiments and Basic Client Refactor by @yc7z in #124
• Adding an implementation for FedDG-GA by @lotif in #117
• Add MR-MTL Method by @sanaAyrml in #125
• Updating the start numpy client code. by @emersodb in #126
• Update tqdm (pip-audit found a vulnerability) by @amrit110 in #133
• Bump jinja2 from 3.1.3 to 3.1.4 by @dependabot in #140
• Bump werkzeug from 3.0.1 to 3.0.3 by @dependabot in #139
• Enable pre-commit ci and add default config by @amrit110 in #132
• Skip flake8 on pre-commit bot by @lotif in #141
• Implementing FedRep by @emersodb in #127
• Adding the ability to specify a checkpointing for models before aggregation. by @emersodb in #128
• Creating an Opacus specific Checkpointer by @emersodb in #129
• Forcing better model consistency for Server-side models with Opacus by @emersodb in #130
• Consolidate Weight Drift Losses by @emersodb in #131
• Add dependabot config to update github actions by @amrit110 in #142
• Bump actions/setup-python from 3 to 5 by @dependabot in #143
• Bump actions/checkout from 3 to 4 by @dependabot in #144
• Bump actions/cache from 2 to 4 by @dependabot in #145
• Bump pypa/gh-action-pypi-publish from 1.4.2 to 1.8.14 by @dependabot in #146

New Contributors

• @dependabot made their first contribution in #140

Full Changelog: v0.1.13...v0.1.15

Add Ditto, Aggregation utils, PICAI dataset utils

What's Changed

• Aggregation utils to main lib by @jewelltaylor in #110
• Picai example by @jewelltaylor in #101
• Implementing Ditto for paper experimentation by @emersodb in #111
• Increment release version by @jewelltaylor in #113

Full Changelog: v0.1.12...v0.1.13

Sparse Parameter Exchanger and Federated Autoencoders.

What's Changed

• Relax Floating Point Number Tolerance for PCA test by @yc7z in #106
• Add separate functions to save aggregated and old local models in Fenda and Moon by @sanaAyrml in #109
• Sparse Parameter Exchange by @yc7z in #103
• Federated ae by @fatemetkl in #82

Full Changelog: v0.1.11...v0.1.12

Smoke Tests, Fed PCA, Small Bug Fixes, README update

What's Changed

• Upgrading jinja2 to 3.1.3 to fix a security issue by @lotif in #96
• Add federated checkpointing by @sanaAyrml in #95
• Federated PCA by @yc7z in #89
• Add a reporter to save the metrics to a parseable file by @lotif in #92
• Smoke tests: Fix random seed on smoke tests and add asserts on results by @lotif in #97
• Ignoring one more pillow 9.5.0 vulnerability by @lotif in #99
• Initial Parameter Retrieval Code Duplication Reduction by @emersodb in #104
• README Update to include all of the new work! by @emersodb in #105
• Splitting Loss into EvaluationLoss and TrainLoss by @lotif in #100
• Update personalized approaches section header by @amrit110 in #107
• Fixing a small, but significant, bug with MOON. by @emersodb in #108

Full Changelog: v0.1.10...v0.1.11