v1.4.4

orangetin released this 12 Mar 19:10

· 21 commits to main since this release

639fb48

What's Changed

Add support for the Training Method for finetuning, and for Direct-Preference Optimization (DPO) by @VProv in #262
Fix from_checkpoint fields by @artek0chumak in #265
Update actions in the PyPI release workflow by @artek0chumak in #267

New Contributors

@VProv made their first contribution in #262

Full Changelog: v1.4.3...v1.4.4

Contributors

artek0chumak and VProv

Assets 2