[SDPA] relax restriction on L from query and attention_mask being equal #24745
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
e-nugmanova
changed the title
e-nugmanova
commented
May 28, 2024
src/core/shape_inference/include/scaled_dot_product_attention_shape_inference.hpp
Outdated
Show resolved
Hide resolved
src/core/shape_inference/include/scaled_dot_product_attention_shape_inference.hpp
Outdated
Show resolved
Hide resolved
mitruska
approved these changes
May 28, 2024
Contributor
mitruska
left a comment
mitruska
left a comment
There was a problem hiding this comment.
LGTM, but would be great to cover this change with corresponding tests.
The SDPA spec clarification can be handled as a follow up.
src/core/shape_inference/include/scaled_dot_product_attention_shape_inference.hpp
Show resolved
Hide resolved
mmikolajcz
approved these changes
May 28, 2024
This was referenced Jun 25, 2024
allnes
pushed a commit
to allnes/openvino
that referenced
this pull request
Jun 27, 2024
…al (openvinotoolkit#24745) ### Details: - *relax restriction on L from query and attention_mask being equal* ### Tickets: - *CVS-129000*
AsyaPronina
pushed a commit
to AsyaPronina/openvino
that referenced
this pull request
Jul 1, 2024
…it#25208) ### Details: - Update attention mask expected shape, allow for broadcast (aligned with shape_infer, decomposition, and CPU plugin) - Fix existing batch broadcast example - Update `scale` input shape info (allow for single element 1D tensor) - related PRs: openvinotoolkit#24745 openvinotoolkit#25213 ### Tickets: - 143009
AsyaPronina
pushed a commit
to AsyaPronina/openvino
that referenced
this pull request
Jul 1, 2024
… input broadcast (openvinotoolkit#25213) ### Details: - More ScaledDotProductAttention tests to cover attention mask input broadcast - related PRs: openvinotoolkit#24745 openvinotoolkit#25213 ### Tickets: - 143009
AsyaPronina
pushed a commit
to AsyaPronina/openvino
that referenced
this pull request
Jul 1, 2024
…it#25208) ### Details: - Update attention mask expected shape, allow for broadcast (aligned with shape_infer, decomposition, and CPU plugin) - Fix existing batch broadcast example - Update `scale` input shape info (allow for single element 1D tensor) - related PRs: openvinotoolkit#24745 openvinotoolkit#25213 ### Tickets: - 143009
AsyaPronina
pushed a commit
to AsyaPronina/openvino
that referenced
this pull request
Jul 1, 2024
… input broadcast (openvinotoolkit#25213) ### Details: - More ScaledDotProductAttention tests to cover attention mask input broadcast - related PRs: openvinotoolkit#24745 openvinotoolkit#25213 ### Tickets: - 143009
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Details:
Tickets: