Fix SFT loss type rewards being overwritten in dpo_loss()#5079
Merged
qgallouedec merged 1 commit intohuggingface:mainfrom Feb 16, 2026
Merged
Fix SFT loss type rewards being overwritten in dpo_loss()#5079qgallouedec merged 1 commit intohuggingface:mainfrom
qgallouedec merged 1 commit intohuggingface:mainfrom