Use fused SDPA for SVTRv2 global attention - #221
Open
kawavil wants to merge 2 commits into
Open
Conversation
kawavil
marked this pull request as draft
September 11, 2026 06:32
kawavil
marked this pull request as ready for review
September 11, 2026 06:32
kawavil
marked this pull request as draft
September 11, 2026 08:50
kawavil
marked this pull request as ready for review
September 11, 2026 08:53
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Use fused SDPA for SVTRv2 global attention to avoid memory issue. During inference, manual attention calculation stores the metrics in the memory and it consumes lot of memory. So to avoid that we can use the pytorch SDPA. Check the comparison below:
Metric | Manual attention | SDPA attention
Test images | 200 | 200
Correct predictions | 187 | 187
Accuracy | 93.50% | 93.50%
Total inference time| 116.2 sec | 42.1 sec
Peak GPU allocated | 4936.21 MB | 245.78 MB
Peak GPU reserved | 6610.00 MB | 330.00 MB