david/flash-attention: flash-attention from https://github.com/Dao-AILab/flash-attention @ tdd

mirrorاز https://github.com/Dao-AILab/flash-attention

Antoni Viros 83e41b3ca4 Add custom ops for compatibility with PT Compile (#1139)		6 ماه پیش
..
src	65f723bb9a Split bwd into more .cu files to speed up compilation	8 ماه پیش
flash_api.cpp	83e41b3ca4 Add custom ops for compatibility with PT Compile (#1139)	6 ماه پیش