david/flash-attention

mirror da https://github.com/Dao-AILab/flash-attention

Autore	SHA1 Messaggio	Data
Tri Dao	82c1aa3514 Move PackGQA epilogue code to pack_gqa.h	2 mesi fa
Tri Dao	df96486c31 Decode: varlen, paged KV, leftpad	2 mesi fa
Ying Zhang	3669b25206 bwd benchmark + small fixes (#1129)	6 mesi fa
Ying Zhang	c7f20a2d31 add cudnn benchmark for var-len	6 mesi fa
Ying Zhang	dfe1a59e4b Add var-seq-len to FA3 fp16 / bf16 fwd (#1072)	6 mesi fa