PR #22292 DecoderMaskedMultiHeadAttention CPU kernel.

DecoderMaskedMultiHeadAttention CPU kernel. #22292

tianleiwu merged 14 commits into main from linmin/cpu_dmmha

DecoderMaskedMultiHeadAttention CPU kernel.

10cd0279

mindest requested a review from

kunal-vaishnavi 1 year ago

github-advanced-security commented on 2024-10-02

Fix attention for no-beam case

f4dd1dc9

github-advanced-security commented on 2024-10-10

Fix errors; update unit test cases

33fd0f10

github-advanced-security commented on 2024-10-10

Fix some CI errors.

0af5deb1

tianleiwu commented on 2024-10-10

Fix: pick up local unstaged changes.

5e745902

tianleiwu commented on 2024-10-11

github-advanced-security commented on 2024-10-11

Fix error; add broadcast for attn_bias; resolve comments

1c90af9f

github-advanced-security commented on 2024-10-11

Format

ae766526

mindest marked this pull request as ready for review 1 year ago

Update doc.

caddf065

Add updated op kernel doc.

b2e35b46

mindest changed the title ~~[WIP] DecoderMaskedMultiHeadAttention CPU kernel.~~ DecoderMaskedMultiHeadAttention CPU kernel. 1 year ago

tianleiwu commented on 2024-10-11

Resolve comments.

4f96ddb4

tianleiwu commented on 2024-10-11

github-advanced-security commented on 2024-10-11

Resolve more comments; fix warning

3ee8d538

Fix CI, warnings.

f9f4ff36

tianleiwu dismissed these changes on 2024-10-12

kunal-vaishnavi commented on 2024-10-12

github-advanced-security commented on 2024-10-12

Fix warning; rename to output_qk

5b5e7913

mindest dismissed their stale review via 5b5e7913 1 year ago

typo

67a653e3

tianleiwu approved these changes on 2024-10-12

tianleiwu merged 1fa219d7 into main 1 year ago

tianleiwu deleted the linmin/cpu_dmmha branch 1 year ago

Reviewers

tianleiwu

kunal-vaishnavi

github-advanced-security

Assignees

No one assigned

Labels

None yet

Milestone

No milestone

onnxruntime DecoderMaskedMultiHeadAttention CPU kernel. #22292 Merged

DecoderMaskedMultiHeadAttention CPU kernel. #22292

onnxruntime
DecoderMaskedMultiHeadAttention CPU kernel.
#22292

Merged