david/flash-attention

Mirror von https://github.com/Dao-AILab/flash-attention

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