DeepSpeed
Fix AutoTP gathering replaced layer params when bias is not None
#7257
Merged

Fix AutoTP gathering replaced layer params when bias is not None #7257

hwchen2017 merged 7 commits into deepspeedai:master from qwen_gether
HollowMan6
HollowMan6 HollowMan6 requested a review from hwchen2017 hwchen2017 1 year ago
HollowMan6 HollowMan6 requested a review from loadams loadams 1 year ago
delock
inkcherry
inkcherry commented on 2025-04-29
HollowMan6 HollowMan6 requested a review from tjruwase tjruwase 1 year ago
HollowMan6 HollowMan6 requested a review from tohtana tohtana 1 year ago
HollowMan6 HollowMan6 changed the title Fix QWen AutoTP when gathering replaced layer params Fix AutoTP gathering replaced layer params when bias is not None 1 year ago
HollowMan6 HollowMan6 requested a review from inkcherry inkcherry 1 year ago
inkcherry
HollowMan6
hwchen2017
hwchen2017 approved these changes on 2025-04-30
loadams
loadams approved these changes on 2025-05-08
HollowMan6
loadams
loadams loadams enabled auto-merge 360 days ago
loadams
disabled auto-merge 359 days ago
Head branch was pushed to by a user without write access
HollowMan6
inkcherry
HollowMan6
HollowMan6 Fix AutoTP gathering replaced layer params when bias is not None
3d6e4c53
hwchen2017 Merge branch 'master' into qwen_gether
67c3e416
inkcherry
inkcherry
inkcherry fix norm
f2137b77
HollowMan6 Merge branch 'master' into qwen_gether
b22d49dd
hwchen2017 Merge branch 'master' into qwen_gether
e1d5d106
HollowMan6
HollowMan6 commented on 2025-05-24
HollowMan6 Fix tests of expected tp params for row-parallel
5c2a283a
inkcherry update
db1c174b
hwchen2017 hwchen2017 merged b666844f into master 354 days ago

Login to write a write a comment.

Login via GitHub

Assignees
No one assigned
Labels
Milestone