Allow GQA to output attention scores #25319
Open
+373
−69
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
1. Add optional output to CPU impl of GQA op for storing attention scores (QK). Buffer is of shape (B, N, S, T) and can either be fp16 or fp32, depending on the type of other inputs 2. Add `output_post_softmax_qk` attribute to GQA, which controls if attention scores should be saved before or after softmax is applied 3. Add unit tests to cover this use case
derdeljan-msft
requested review from
tianleiwu,
jywu-msft,
kunal-vaishnavi and
aciddelgado
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Description
output_post_softmax_qkattribute to GQA, which controls if attention scores should be saved before or after softmax is applied