verl-project/verl

Support Generative Reward Model (GenRM)

クローズ

#229 opened on 2025/02/09

 (17 件のコメント) (0 件のリアクション) (1 人の担当者)Python (4,427 件のフォーク)auto 404
enhancementgood first issue

Repository metrics

Stars
 (23,038 個のスター)
PR merge metrics
 (平均マージ 3d 6h) (30d で 125 merged PRs)

説明

According to the documentation, veRL only supports AutoModelForSequenceClassification. What would be the best way to implement generative reward model (GenRM) for veRL? I tried looking at FSDP Workers and Megatron-LM Workers but they no longer exist.

コントリビューターガイド