david/flash-attention

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

Autor	SHA1 Správa	Dátum
Tri Dao	dfe29f5e2b [Gen] Don't use ft_attention, use flash_attn_with_kvcache instead	1 rok pred
Tri Dao	8a733cbd53 [Gen] Fix calling update_graph_cache in tests	1 rok pred
Tri Dao	913922cac5 [Gen] Refactor decoding function	1 rok pred
Tri Dao	798858f9f1 Fix test_baichuan	1 rok pred
Tri Dao	0e8c46ae08 Run isort and black on test files	1 rok pred
Tri Dao	184b992dcb [GPT] Implement parallel LLaMa	1 rok pred
Tri Dao	d38357dd2f [GPT] Implement Falcon	1 rok pred