Code release - Githubissues

huggingface / alignment-handbook

Robust recipes to align language models with human and AI preferences

https://huggingface.co/HuggingFaceH4

Apache License 2.0

4.6k stars 401 forks source link

Code release #11

Closed lewtun closed 11 months ago

lewtun commented 11 months ago

TODO

[x] Add instructions on how to train Zephyr
[x] Add LoRA configs and QLoRA configs
[x] Evaluate Zephyr SFT & DPO models with MT-Bench
[x] Add unit tests or YOLO?

Closes #6