Releases: artefactory/choice-learn
Releases · artefactory/choice-learn
v1.3.2 Changes in dependencies
What's Changed
- ADD: model get-set weights by @VincentAuriau in #301
- Change version to v1.3.1 by @VincentAuriau in #298
- Fix bad data by @VincentAuriau in #311
- Fix bad data by @VincentAuriau in #312
- quick documentation fix by @VincentAuriau in #313
- New batching by @michaelteboul in #308
- ENH: upgraded allowed TF versions by @VincentAuriau in #316
- Update tf version by @VincentAuriau in #317
Full Changelog: v1.3.1...v1.3.2
v1.3.1: fixed early stopping & add possibility for weighted validation dataset
What's Changed
- Fix fit2 by @michaelteboul in #293
- ENH: early stopping in models & example with RUMnet by @VincentAuriau in #294
- ADD: samples weight for val_dataset in .fit & validation_freq arg by @VincentAuriau in #295
- Change version to v1.3.0 by @VincentAuriau in #297
Full Changelog: v1.3.0...v1.3.1
ADD: self-Attention model, Bakery Dataset & more
What's Changed
- Change version to v1.2.0 by @VincentAuriau in #249
- Small changes by @julesdesir in #252
- ADD: latest documentation by @VincentAuriau in #251
- Add base basket class by @VincentAuriau in #256
- ADD: & FIX different docs by @VincentAuriau in #257
- ADD: Embedding model w/ basic attention by @VincentAuriau in #253
- ADD: method & test for cLogit save by @VincentAuriau in #260
- Fix AleaCarta variable names by @VincentAuriau in #261
- FIX: wrong test by @VincentAuriau in #264
- Correctifs sur base.py et notebooks by @michaelteboul in #262
- Update README.md by @EmmanuelMalherbe in #263
- [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in #241
- Add a new author to allow comments on pull request by @michaelteboul in #267
- London dataset by @michaelteboul in #265
- ADD: better Sample weight handling by @VincentAuriau in #269
- Bakery dataset by @michaelteboul in #266
- ENH: Bakery Dataset by @VincentAuriau in #272
- ADD: Official Citation version by @VincentAuriau in #273
- ADD: optimizer state can now be saved & reinstantiated by @VincentAuriau in #279
- update & fix notebook with recent changes by @VincentAuriau in #276
- FIX: evaluate + few enh by @VincentAuriau in #280
- Enh eval by @VincentAuriau in #283
- Some doc by @VincentAuriau in #285
- ENH: Rm - small edits by @VincentAuriau in #288
- Add batch eval by @VincentAuriau in #287
- ADD: self-attention model by @michaelteboul in #281
- Fix evaluate in fit by @michaelteboul in #290
- [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in #270
New Contributors
- @michaelteboul made their first contribution in #262
Full Changelog: v1.2.0...v1.3.0
v1.2.0 Introduction of basket models !
New features
This new version mainly introduces two new discrete choice models the Halo-MNL as well as its low-rank version.
Basket choice models !
A brand new type of models is also integrated: basket choice models.
The first two models added are Shopper and the AleaCarta models.
As always references can be found in the ReadMe.
Basket choice will be one of the main focus for the future additions, stay tuned !
Improvements
Diverse fixes and improvement are also released
Detailed changements:
- Change version to v1.1.1 by @VincentAuriau in #192
- [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in #195
- [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in #198
- ADD: Latent Class Report by @VincentAuriau in #201
- Fix test report by @VincentAuriau in #203
- [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in #204
- ENH: Expectation-Maximization Algorithm by @VincentAuriau in #205
- Add lc endpoint: predict_modelwise_probas by @VincentAuriau in #206
- ADD: Data tests by @VincentAuriau in #207
- ADD: Tools tests by @VincentAuriau in #208
- ADD: Model tests by @VincentAuriau in #209
- FIX: CI doublons by @VincentAuriau in #210
- [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in #211
- ADD: basic Halo-MNL and LowRank-Halo-MNL by @VincentAuriau in #213
- FIX: full dataset indexation when 1D items features by @VincentAuriau in #214
- [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in #215
- [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in #218
- Few changes by @VincentAuriau in #220
- Basket models by @julesdesir in #216
- Fix documentation by @julesdesir in #226
- FIX: diverse elements of documentation by @VincentAuriau in #227
- [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in #219
- Fix doc by @VincentAuriau in #228
- Fix doc by @VincentAuriau in #229
- FIX: small typo by @julesdesir in #230
- ENH: batch processing for Shopper.evaluate() by @julesdesir in #231
- ENH: no trainable lambda for the checkout item in Shopper by @julesdesir in #232
- Increase performance with tf.function decorator by @julesdesir in #234
- [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in #233
- FIX: Pathlib use in load_model by @VincentAuriau in #238
- [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in #237
- Update PR comment by @VincentAuriau in #242
- FIX: doc by @VincentAuriau in #244
- ADD: ECML-PKDD model by @VincentAuriau in #245
Full Changelog: v1.1.1...v1.2.1
FIX: tolerance argument for Latent Class model
What's Changed
- Change version to v1.1.0 by @VincentAuriau in #181
- [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in #178
- [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in #185
- Update actions to satisfy minimum nodejs version by @chicham in #187
- [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in #188
- Fix tolerance argument in LatentClass Model by @VincentAuriau in #190
- ADD: few simple LC model tests by @VincentAuriau in #191
New Contributors
- @pre-commit-ci made their first contribution in #178
Full Changelog: v1.1.0...v1.1.1
v1.1.0: Learning-MNL, ResLogit & London Passenger
Main changes:
- ADD ResLogit and Learning-MNL (Enhanced DCM) models
- ADD London Passenger dataset
- ADD: possible parametrization of nb of parallel iterations when usign lbfgs
- ADD released paper & citation
- Diverse small improvements and fixes
Details:
- Release v1.0.0 by @VincentAuriau in #156
- ADD: official paper & citation by @VincentAuriau in #158
- ADD: London Passenger Choice Mode by @VincentAuriau in #159
- ADD: tests for SimpleMNL and ChoiceModel by @VincentAuriau in #160
- Test eager mode for some model tests by @VincentAuriau in #161
- microFIX: NestedLogit clone by @VincentAuriau in #163
- Add ResLogit implementation and a notebook example by @julesdesir in #162
- Minor enhancements by @VincentAuriau in #164
- TST: fix test_reslogit.py by @julesdesir in #165
- TST: Switch to "optim" mode in model.evaluate() by @julesdesir in #166
- ADD: allow for choice of nb of parallel iterations when using LBFGS by @VincentAuriau in #167
- ADD: diverse tests for models by @VincentAuriau in #168
- Add RUMnet & Nested Logit tests by @VincentAuriau in #169
- ADD: White logo by @VincentAuriau in #170
- ADD: Learning-MNL by @VincentAuriau in #171
- ENH: README by @VincentAuriau in #173
- Add colab notebooks links by @VincentAuriau in #175
- Make hooks faster by @chicham in #177
New Contributors
- @julesdesir made their first contribution in #162
- @chicham made their first contribution in #175
Full Changelog: v1.0.0...v1.1.0
✨✨ Release of v1.0.0 ✨✨
✨✨ Release of v1.0.0 ✨✨
What's Changed
- Fix EM algorithm to fit Latent Class model by @VincentAuriau in #152
- ADD: JoSS submission Paper by @VincentAuriau in #155
Full Changelog: v0.0.8...v1.0.0
v0.0.8: FIX: tensorflow-probability version
v0.0.7: Wider allowed dependencies
What's Changed
- ENH: broader poetry dependencies by @VincentAuriau in #139
- Fix dependencies by @VincentAuriau in #141
Full Changelog: v0.0.6...v0.0.7
v0.0.6
What's Changed
- ENH: default n_epochs for Latent models is now 100 by @VincentAuriau in #122
- FIX: RUMnet regularization & evaluate by @VincentAuriau in #126
- FIX: ReadMe example by @VincentAuriau in #125
- ENH: requirements.txt allow for broader versions of dependencies by @VincentAuriau in #134
- ENH: ChoiceModel.trainable_weights are now defined with decorator property by @VincentAuriau in #116
Full Changelog: v0.0.5...v0.0.6