Add something that cannot be achieved before.


Improve the efficiency on the computation or memory.


Miscellaneous minor improvements.


Fix up something that does not work as expected.

API Change

You will need to change the code to have the same effect.

Ver 0.1.*

  • Feature API Change Support using dataloader with multiple input | @xuyxu

  • Fix Fix missing functionality of use_reduction_sum for fit() of Gradient Boosting | @xuyxu

  • Enhancement Relax tensorboard as a soft dependency | @xuyxu

  • Enhancement API Change Simplify the training workflow of FastGeometricClassifier and FastGeometricRegressor | @xuyxu

  • Feature API Change Support TensorBoard logging in set_logger() | @zzzzwj

  • Enhancement API Change Add use_reduction_sum parameter for fit() of Gradient Boosting | @xuyxu

  • Feature API Change Improve the functionality of evaluate() and predict() | @xuyxu

  • Feature API Change Add FastGeometricClassifier and FastGeometricRegressor | @xuyxu

  • Enhancement Add flexible instantiation of optimizers and schedulers | @cspsampedro

  • Feature API Change Add support on accepting instantiated base estimators as valid input | @xuyxu

  • Fix Fix missing base estimators when calling load() for all ensembles | @xuyxu

  • Feature API Change Add methods on model deserialization load() for all ensembles | @mttgdd


  • Feature API Change Add set_scheduler() for all ensembles | @xuyxu

  • Major Feature Add AdversarialTrainingClassifier and AdversarialTrainingRegressor | @xuyxu

  • Major Feature Add SnapshotEnsembleClassifier and SnapshotEnsembleRegressor | @xuyxu

  • Feature API Change Add model validation and serialization | @ozanpkr and @xuyxu

  • Enhancement Add CI and maintenance tools | @xuyxu

  • Enhancement Add the code coverage on codecov | @xuyxu

  • Enhancement Add the version numbers to requirements.txt | @zackhardtoname and @xuyxu

  • Enhancement Improve the logging module using logging | @zzzzwj

  • API Change Remove the input argument output_dim from all methods | @xuyxu

  • API Change Refactor the setup on optimizer into set_optimizer() | @xuyxu

  • API Change Refactor the codes on operating tensors into an independent module | @zzzzwj

  • Fix Fix the bug in logging module when using multi-processing | @zzzzwj

  • Fix Fix the binding problem on scheduler and optimizer when using parallelization | @Alex-Medium and @xuyxu