david/flash-attention

-ын хуулбар https://github.com/Dao-AILab/flash-attention

Эзэн	SHA1 Мессеж	Огноо
Jay Shah	5e3864f2ee change default output type of fp8 kernel to bf16	2 сар өмнө
Jay Shah	b5cac6d586 rebase with Is_local disabled temporarily	2 сар өмнө
jayhshah	c92ca63268 FA3 FP8 qkv descales + restore max offset for h128 causal + added sync for producer WG (#1173)	3 сар өмнө
jayhshah	5018ac6ac5 Fp8 kernel with "in-kernel" transpose of V in producer (#1100)	4 сар өмнө
Tri Dao	3aae9c18c1 Revert "Changes For FP8 (#1075)"	4 сар өмнө
ganeshcolfax	1899c970c8 Changes For FP8 (#1075)	4 сар өмнө