DeepSpeed
Fix repeated gradient scaling across pipeline stages
#8154
Merged

Fix repeated gradient scaling across pipeline stages #8154

Functionhx
Functionhx Fix pipeline gradient accumulation scaling
1cc1c820
Functionhx Functionhx marked this pull request as ready for review 29 days ago
Functionhx Functionhx requested a review from loadams loadams 29 days ago
Functionhx Functionhx requested a review from tohtana tohtana 29 days ago
Functionhx Functionhx requested a review from tjruwase tjruwase 29 days ago
tohtana
tohtana approved these changes on 2026-07-20
tohtana tohtana enabled auto-merge 26 days ago
PKUWZP
PKUWZP approved these changes on 2026-07-20
tohtana tohtana merged d99a6429 into master 26 days ago

Login to write a write a comment.

Login via GitHub

Assignees
No one assigned
Labels
Milestone