PR #16743 Fix transformers optimizations for GPT-NeoX

Fix transformers optimizations for GPT-NeoX #16743

tianleiwu merged 12 commits into main from tlwu/fix_fp16_model_optimizer

Fix issue for GPT-NeoX

18dc137d

tianleiwu requested a review from

wangyems 2 years ago

simplify use_float16

3aa37002

remove unused import

a870fffb

sort models by name

0320020c

refine

3780a7b4

change default opt_level to 0

e1d26ec7

change a log level to debug

97377b7a

update log message

b0ea8c18

fix typo

715d1f18

Do not load external data for checking fp16

7d9284cc

update comment

054d6347

yufenglee commented on 2023-07-18

yufenglee dismissed these changes on 2023-07-18

check attribute type

7fe64fff

tianleiwu dismissed their stale review via 7fe64fff 2 years ago

yufenglee approved these changes on 2023-07-18

tianleiwu merged 95f96287 into main 2 years ago

tianleiwu deleted the tlwu/fix_fp16_model_optimizer branch 2 years ago

Reviewers

yufenglee

wangyems

Assignees

No one assigned

Labels

None yet

Milestone

No milestone