optuna 0.10.0

v0.10.0

on Python PyPI

This is the release note of v0.10.0. See here for the complete list of solved issues and merged PRs.

Enhancements

• Add GC call after evaluating an objective function. (#377)
• Change order of checking code. (#374, thanks @okdshin!)
• Restrict data load to worker 0. (#362)
• Add cache to RDBStorage. (#349)
• Inhibit storage access of samplers in ChainerMNStudy. (#348)
• Introduce Alembic. (#336)

Bug Fixes

• Fix minor problems of TPESampler and ParzenEstimator. (#373)
• Fix an issue in parallel execution where inconsistent FronzenTrial could be created with low probability. (#361)

Documents

• Add an issue template. (#378, thanks @crcrpar!)
• Add KerasPruningCallback and TensorFlowPruningHook to document. (#371)
• Rewrite formulas using mathjax. (#367, thanks @crcrpar!)

Examples

• Add Keras pruning example. (#364, thanks @Rishav1!)
• Add simple MXNet example for MNIST. (#359, thanks @Rishav1!)

Tests

• Enable tensorflow in Python 3.7. (#360)