transformers
Fix initialization for missing parameters in `from_pretrained` under ZeRO-3
#28245
Merged

Fix initialization for missing parameters in `from_pretrained` under ZeRO-3 #28245

XuehaiPan
XuehaiPan Fix initialization for missing parameters in `from_pretrained` under …
ceda28b3
pacman100
pacman100 commented on 2023-12-27
pacman100 pacman100 requested a review from amyeroberts amyeroberts 2 years ago
pacman100 pacman100 requested a review from ArthurZucker ArthurZucker 2 years ago
XuehaiPan Test initialization for missing parameters under ZeRO-3
33e67e06
XuehaiPan XuehaiPan force pushed to 33e67e06 2 years ago
XuehaiPan XuehaiPan requested a review from pacman100 pacman100 2 years ago
XuehaiPan
XuehaiPan Add more tests
10ae1845
ArthurZucker
ArthurZucker commented on 2024-01-03
XuehaiPan Only enable deepspeed context for per-module level parameters
855db0fd
XuehaiPan Enable deepspeed context only once
a781cb10
XuehaiPan XuehaiPan requested a review from ArthurZucker ArthurZucker 2 years ago
pacman100
pacman100 approved these changes on 2024-01-08
ArthurZucker
ArthurZucker approved these changes on 2024-01-09
amyeroberts
amyeroberts approved these changes on 2024-01-09
XuehaiPan Move class definition inside test case body
5170a013
XuehaiPan XuehaiPan force pushed to 5170a013 2 years ago
HuggingFaceDocBuilderDev
amyeroberts amyeroberts merged 976189a6 into main 2 years ago
XuehaiPan XuehaiPan deleted the fix-zero-3-init-missing branch 2 years ago

Login to write a write a comment.

Login via GitHub

Assignees
No one assigned
Labels
Milestone