Merge pull request #29126 from abhishek-gola:flash_attention

mirror of https://github.com/opencv/opencv.git synced 2026-07-30 15:53:03 +04:00

Attention graph fusion with MLAS FlashAttention #29126

Performance numbers for Owl-v2 model on intel i9:
```
ORT: Average inference time over 10 runs: 1411.55 ms (min 1399.75, max 1438.89)
NEW: Average inference time over 10 runs: 1078 ms (min 1048.04, max 1110.61)
```
### Pull Request Readiness Checklist

See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request

- [x] I agree to contribute to the project under Apache 2 License.
- [x] To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV
- [x] The PR is proposed to the proper branch
- [x] There is a reference to the original bug report and related work
- [x] There is accuracy test, performance test and test data in opencv_extra repository, if applicable
      Patch to opencv_extra has the same branch name.
- [x] The feature is well documented and sample code can be built with the project CMake

This commit is contained in:

Abhishek Gola

2026-05-27 12:02:27 +05:30

committed by

GitHub

parent 2405556989

commit bf0cf34963

11 changed files with 2049 additions and 41 deletions

									
										3rdparty/mlas/lib/mlasi.h
									
Vendored

		+1
		-1
	
												View File
												
				@@ -408,7 +408,7 @@ size_t

				#if defined(__aarch64__) && defined(__linux__)

				typedef size_t(MLASCALL MLAS_SBGEMM_FLOAT_KERNEL)(

				    const float* A,

				    const bfloat16_t* B,

				    const void* B,

				    float* C,

				    size_t CountK,

				    size_t CountM,