david/flash-attention

peilaus alkaen https://github.com/Dao-AILab/flash-attention

Tekijä	SHA1 Viesti	Päivämäärä
Tri Dao	6e8b25e426 Refactor	2 kuukautta sitten
jayhshah	c92ca63268 FA3 FP8 qkv descales + restore max offset for h128 causal + added sync for producer WG (#1173)	3 kuukautta sitten
Tri Dao	bafe253042 [FA3] Bwd	4 kuukautta sitten
jayhshah	5018ac6ac5 Fp8 kernel with "in-kernel" transpose of V in producer (#1100)	4 kuukautta sitten
Tri Dao	7f67966cc7 FA3 initial code release	5 kuukautta sitten