v0.1.8

What's Changed

Document

• Add Grid Search with Sklearn Estimator
• Generate quickstart by sphinx gallery

Highlights

• API change: move linear.sklearn_helper.* to linear.*

Contributors

Thank you all for your contributions :)

@cjlin1 @will945945945 @Gordon119 @Sinacam @Eleven1Liu

Original release date: Feb 8, 2023

v0.1.7

What's Changed

Highlights

• Add linear solver: binary and multiclass
• Add multi-label estimator
• Add log collecting mechanism
• Load best model without retraining after parameter selection

Contributors

Thank you all for your contributions :)

@cjlin1 @JamesLYC88 @Sinacam @Gordon119 @sammer1107 @Eleven1Liu

New Contributors

Welcome @Gordon119 🎉

Original release date: Jan 21, 2023

v0.1.6

What's Changed

Document

• Added BERT quickstart. See the latest document: https://www.csie.ntu.edu.tw/~cjlin/libmultilabel/.

Highlights

• Enabled early_stopping_metric and Loss metric
• Set multi-class classification automatically

Contributors

Thank you all for your contributions :)

@cjlin1 @Eleven1Liu @sian-chen @JamesLYC88 @Sinacam @will945945945

Original release date: Nov 23, 2022 by @JamesLYC88

v0.1.5

What's Changed

Document

• Added the Command Line Options section in documentation. See the latest document: https://www.csie.ntu.edu.tw/~cjlin/libmultilabel/.

Highlights

• Added support for binary and multi-class classification
• Added re-training with optimal-epochs criterion
• Updated PyTorch version to 1.12
• Added BERT-BASE
• Added a fast basic test on example configs

Contributors
Thank you all for your contributions :)

@cjlin1 @Eleven1Liu @sian-chen @JamesLYC88 @Sinacam @will945945945

Original release date: Sep 30, 2022 by @JamesLYC88

v0.1.3

What's Changed

Document

• Added the User Guide link in documentation. See the latest document: https://www.csie.ntu.edu.tw/~cjlin/libmultilabel/.

Highlights

• Added cost sensitive that optimizes Micro-F1
• Added label and macro options for linear
• Refactored nn.networks
• Added BERT-LWAN and multihead attention

Contributors
Thank you all for your contributions :)

@cjlin1 @Sinacam @ericliu8168 @sian-chen @Eleven1Liu

Original release date: Feb 20, 2022

v0.1.2

What's Changed

Document

• Update installation instructions. See the latest document: https://www.csie.ntu.edu.tw/~cjlin/libmultilabel/.

Highlights

• Decouple linear from torch libraries
• Added linear tests and save k prediction
• Fix Macro-F1 score

Contributors
Thank you all for your contributions :)

@cjlin1 @Sinacam @Eleven1Liu @ericliu8168 @sian-chen

Original release date: Dec 21, 2021

v0.1.1

What's Changed

Document

• Reorganized the document and README. See the latest document: https://www.csie.ntu.edu.tw/~cjlin/libmultilabel/.

Highlights

• Added new functions: thresholding and cost-sensitive to linear classifier
• Support ASHA scheduler in parameter search
• Allow NN trainer to train without the validation set
• Added auto-testing script
• Automated the release process with Github Action

Contributors
Thank you all for your contributions :)

@cjlin1 @Eleven1Liu @Sinacam @ericliu8168 @sian-chen

Original release date: Dec 10, 2021

v0.1.0

What's Changed

• First release of LibMultiLabel (2021.04.14~2021.10.23)
• Documentation: https://www.csie.ntu.edu.tw/~cjlin/libmultilabel/

Contributors

Thank you all for your contributions :)

@cjlin1 @Eleven1Liu @sian-chen @henryyang42 @Sinacam @chihming @optchliu

Original release date: Oct 23, 2021