david/flash-attention

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

نویسنده	SHA1 پیام	تاریخ
Tri Dao	2c996ca25f Use SeqlenInfo for bwd and epilogue	1 هفته پیش
Tri Dao	29cdfedd80 Use Bulk reduce instead of TMA for dQaccum, split across WGs	1 هفته پیش
Tri Dao	82dc825759 Don't use the unsafe convert_type function	2 هفته پیش
Tri Dao	6e8b25e426 Refactor	2 ماه پیش
Ying Zhang	db80387343 Add seqused_q in fwd / bwd and seqused_k in bwd.	3 ماه پیش
Cameron Shinn	3cea2fb6ee Add ArchTag to pre/postprocess bwd kernels (#1180)	3 ماه پیش
Tri Dao	bafe253042 [FA3] Bwd	4 ماه پیش