v0.4.0-rc1

Starting with this release, we will use release candidates! These bring you all the new experimental features from nightly, while including hotfixes from main. We will try to make the new branch release-candidate stable, meaning merges from nightly only happen when things are looking good.

In this release, you may find:

• Improved HParams logging
• Improved default HParams and HParams handling
• A convenient way to load your best model with Trainable.load_best_checkpoint()
• Bugfixes and more!

v0.3.1

This release introduces several hotfixes, among them nested AttributeDicts, dict-type HParams, as well as LRScheduler configs.

v0.3.0

This release brings you modules, metrics, bugfixes and more!

v0.2.0

This release brings you Modules, Metrics & More!

v0.0.3

This release brings you experiment launchers!

v0.0.2

version bump

v0.0.1

Initial Release