optuna 0.17.0

v0.17.0

on Python PyPI

This is the release note of v0.17.0. See here for the complete list of solved issues and merged PRs.

Dropping Support of Python 2

Due to the end-of-life (EOL) of Python 2 in January 2020, Optuna will drop Python 2 support in December 2019.

This decision was made considering the following facts:

• Python 2 will become end-of-life (EOL) in January 2020.
• Many scientific computation packages, including NumPy, which is one of the core dependency of Optuna, are planning or already started to drop support for Python 2.

We plan to drop Python 2 support in December 2019, but the date has not been determined yet. The detailed schedule will be published in the next release.

New Features

• Add NopPruner. (#555, thanks @Muragaruhae)
• Add pruning callback for tf.keras. (#530, thanks @sfujiwara!)
• Add delete_study API for storage package. (#524, thanks @c-bata!)
• Add function for optimization history visualization. (#513, thanks @suecharo!)
• Add user-defined callbacks. (#480)
• Add Study._append_trial method. (#464)
• Add OptunaSearchCV that provides sklearn compatible API (experimental). (#357, thanks @Y-oHr-N!)

Enhancements

• Add maximize support at dashboard. (#587, thanks @c-bata!)
• Make study returns an empty dataframe if no trials. (#574, thanks @crcrpar!)
• Support TensorFlow 2.0. (#562)
• Add a function to delete a study object. (#558, thanks @c-bata!)
• Fix type hints for mypy==0.730. (#553)
• Add datetime_start property to FixedTrial. (#548)
• Update high value using Decimal. (#546)
• Apply small fixes to history and intermediate plots. (#541, thanks @suecharo!)
• Fix test_keras_pruning_callback to be compatible with keras==2.3.0. (#537)
• Change the deprecated product_search_space to intersection_search_space. (#532, thanks @oda!)
• Add datetime_start to ChainerMN integration. (#528, thanks @tanapiyo!)
• Remove dependency on pandas. (#527, thanks @henry0312!)
• Add Python3.4 version check to RDBStorage. (#525, thanks @tadan18!)
• Add datetime_start during a trial. (#523, thanks @tanapiyo!)
• Move example's setup to setup.py. (#522, thanks @tanapiyo!)
• Change class to use for type checking. (#519, thanks @suecharo!)
• Support plotly version 4.0.0. (#511, thanks @suecharo!)
• Fix round-off errors by using decimal module #503. (#505, thanks @A03ki!)
• Add cascade settings. (#422, thanks @c-bata!)

Bug Fixes

• Fix package of TracebackType. (#547)

Documents

• Remove init arguments from study docs. (#579, thanks @scouvreur!)
• Update explanation of the effect of round-off errors on suggest_discrete_uniform. (#544)

Examples

• Update .gitignore to ignore files related to examples. (#578, thanks @crcrpar!)
• Make LightGBM examples more practical. (#575)
• Add an example of MXNetPruningCallback. (#570, thanks @crcrpar!)
• Improve clarity of MXNet example. (#569, thanks @crcrpar!)
• Add PyTorch Ignite example. (#559)
• Fix OOM errors during execution of TF Keras example on CircleCI. (#545)
• Apply maximize direction. (#526, thanks @r-soga!)
• Simplify LightGBM examples. (#580)