feat: expose pytorch api for block sparse attention #375
Conversation
This file contains 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
yzh119
pushed a commit
that referenced
this pull request
Jul 17, 2024
🤖 I have created a release *beep* *boop* --- ## [0.1.0](v0.0.9...v0.1.0) (2024-07-17) ### Features * Add mask to `merge_state_in_place` ([#372](#372)) ([e14fa81](e14fa81)) * expose pytorch api for block sparse attention ([#375](#375)) ([4bba6fa](4bba6fa)) * Fused GPU sampling kernel for joint top-k & top-p sampling ([#374](#374)) ([6e028eb](6e028eb)) --- This PR was generated with [Release Please](https://github.com/googleapis/release-please). See [documentation](https://github.com/googleapis/release-please#release-please). Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
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Outdated suggestions cannot be applied.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.
The block sparse attention (for any block size (R, C)) are hidden in flashinfer's codebase but it was never exposed explicitly in python. As requested in #367 , this PR implements the PyTorch APIs for block sparse attention, accordingly to our experiments, it can greatly accelerate attention computation with low density (10x for Tree Attention in Sequoia).