[Feature] reward model inferencer and dpov2 aligner#867

Merged

research4pan merged 14 commits intomainfrom

yizhenjia-dpo-v2

Jun 30, 2024

Commits on Jun 23, 2024

[Feature] add conversation template arg for rm infer
wheresmyhair
committed
[Usability] better to set overwrite_cache for rm inference in online rlhf
wheresmyhair
committed

Commits on Jun 24, 2024

[Feature] Add dpo support
wheresmyhair
committed

Commits on Jun 25, 2024

[Usability] vllm inferencer save in utf-8
wheresmyhair
committed

Commits on Jun 26, 2024

Commits on Jun 27, 2024

Commits on Jun 28, 2024

[Usability] specify trl version
wheresmyhair
committed