[hotfix] fix entropy calculate for pp; using old logp to calculate loss#6364
Merged
TongLi3701 merged 1 commit intogrpo-latestfrom Jul 22, 2025
Merged
[hotfix] fix entropy calculate for pp; using old logp to calculate loss#6364TongLi3701 merged 1 commit intogrpo-latestfrom
TongLi3701 merged 1 commit intogrpo-latestfrom
Commits
Commits on Jul 21, 2025
- committed