PR #24079 [WebNN EP] Support MultiHeadAttention(MHA)

[WebNN EP] Support MultiHeadAttention(MHA) #24079

fdwr merged 9 commits into microsoft:main from peishenyan:mha_attention

peishenyan marked this pull request as draft 1 year ago

An implementation for MHA

b4e43fb1

fix concat past kv bugs

7d2de767

fix mha output cast bugs

cc1b9643

solve the conflict after code rebase

09de3596

update comments

8a50801e

peishenyan force pushed from 5e67f9b1 to 09de3596 1 year ago

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

Honry commented on 2025-04-11

fdwr commented on 2025-04-11

address comments

cfe51a70

format document

75f2b0e5

fdwr commented on 2025-04-13

remove some unnecessary code

bb8815cf

fdwr dismissed these changes on 2025-04-14

add MHA to webnn-operators.md

01659e57

peishenyan dismissed their stale review via 01659e57 1 year ago

fdwr approved these changes on 2025-04-14

Honry approved these changes on 2025-04-14

fdwr merged df581e1c into main 1 year ago

Reviewers

fdwr

Honry

Assignees

No one assigned

Labels

None yet

Milestone

No milestone