[feat](PA): support pa_decode_gluon and refactor attention ops #42
+274
−141
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
valarLip
reviewed
Dec 17, 2025
| output_zeros=False, | ||
| ) | ||
| else: | ||
| if self.rotary_emb is not None: |
Contributor
Author
There was a problem hiding this comment.
yea, now we have model using reshape_and_cache_with_pertoken_quant ? if none, we don`t wanna introduce new dispatch or if else, if necessary, then we add it.
| self.reduce_indptr = reduce_indptr | ||
| self.reduce_final_map = reduce_final_map | ||
| self.reduce_partial_map = reduce_partial_map | ||
| if block_tables_converted is not None: |
Contributor
Author
There was a problem hiding this comment.
just someone deleted it when rebase main, I will recover it.
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Outdated suggestions cannot be applied.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.
This PR integrates new paged attention triton kernel, it supports
sliding_windowandsinkparams. This PR also refactors attention layers with attention backend dispatch.the accuracy of gpt-oss in gsm8k dataset:
