david/flash-attention

réplica de https://github.com/Dao-AILab/flash-attention

Autor	SHA1 Mensaxe	Data
Ying Zhang	8cbc8a042f small fixes	hai 3 meses
Ying Zhang	cdbbe844b1 minor changes to unpad_input test util func	hai 3 meses
Ying Zhang	db80387343 Add seqused_q in fwd / bwd and seqused_k in bwd.	hai 3 meses
Jay Shah	32792d37ec add missing if condition for key_padding_mask in test_util.py	hai 3 meses
Ying Zhang	53537da422 add a unittest	hai 4 meses
Ying Zhang	dfe1a59e4b Add var-seq-len to FA3 fp16 / bf16 fwd (#1072)	hai 4 meses