onnxruntime
CUDA BFloat16 Refactor
#10085
Merged

CUDA BFloat16 Refactor #10085

Lafi7e merged 10 commits into master from weicwang/bfloat16
Lafi7e
Lafi7e Use BFLoat16 instead of nv_bfloat16 for CUDA
2b3f8f7d
mindest fix BFloat16 rocm
8815c589
Lafi7e remove gemm ut from rocm
5467e8e4
Lafi7e remove reducesum ut from rocm
2cfec118
Lafi7e add biasgelu
2b8868ee
Lafi7e add support for more ops
7dc37da4
Lafi7e Lafi7e added training
Lafi7e Lafi7e requested a review from SherlockNoMad SherlockNoMad 4 years ago
Lafi7e Lafi7e requested a review from weixingzhang weixingzhang 4 years ago
Lafi7e Merge branch 'master' into weicwang/bfloat16
d071935a
Lafi7e fix build error
a3c394af
Lafi7e op md update
882cb356
weixingzhang
weixingzhang commented on 2022-01-06
Lafi7e Merge branch 'master' into weicwang/bfloat16
8c3240dc
weixingzhang
weixingzhang approved these changes on 2022-01-14
Lafi7e Lafi7e merged 44e2db93 into master 4 years ago
Lafi7e Lafi7e deleted the weicwang/bfloat16 branch 4 years ago

Login to write a write a comment.

Login via GitHub

Assignees
No one assigned
Labels
Milestone