Repository navigation
[Fix] attention: Move torch_native/flex_attention CUDA graph disable after platform fallback - #43358
Open
sunshinejiali wants to merge 1 commit into
Open
sunshinejiali wants to merge 1 commit into
sunshinejiali wants to merge 1 commit into
Conversation
sunshinejiali
force-pushed
the
fix/cuda-graph-fallback-order
branch
from
October 9, 2026 10:46
aaa6fe0 to
1f71424
Compare
Author
|
/tag-and-rerun-ci |
sunshinejiali
force-pushed
the
fix/cuda-graph-fallback-order
branch
from
October 20, 2026 09:36
1f71424 to
476468a
Compare
This comment has been minimized.
This comment has been minimized.
Author
|
/tag-and-rerun-ci |
Author
|
Hi maintainers, could someone help assign an appropriate reviewer for this PR? It modifies |
This was referenced Nov 3, 2026
This branch has not been deployed
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Motivation
torch_nativeattention when the platform fallback selects it #43142.Modifications
Added TestAttentionBackendCudaGraph with 3 regression tests:
Tests use _create_model_dir() to avoid the model_path="dummy" early return, and @override_platform(is_cuda=False, has_amx=False) to simulate CPU without AMX.
Accuracy Tests
N/A
Speed Tests and Profiling
N/A
Checklist
CI States
Latest PR Test (Base): ❌ Run #39401972727
Latest PR Test (Extra): ❌ Run #39401972633
Latest PR Test (AMD ROCm 10): ❌ Run #39401972844