david/flash-attention

cermin dari https://github.com/Dao-AILab/flash-attention

Pembuat	SHA1 Pesan	Tanggal
Tri Dao	6e8b25e426 Refactor	3 bulan lalu
Ying Zhang	db80387343 Add seqused_q in fwd / bwd and seqused_k in bwd.	4 bulan lalu
Cameron Shinn	3cea2fb6ee Add ArchTag to pre/postprocess bwd kernels (#1180)	4 bulan lalu
Tri Dao	bafe253042 [FA3] Bwd	5 bulan lalu
Tri Dao	7f67966cc7 FA3 initial code release	5 bulan lalu