Accueil Explorer Aide
S'inscrire Connexion
david
/
flash-attention
miroir de https://github.com/Dao-AILab/flash-attention
1
0
Fork 0
Fichiers Tickets 0 Wiki
Branche: changes_for_fp8
Branches Tags
flash-attention
/
csrc
Tri Dao 65f723bb9a Split bwd into more .cu files to speed up compilation il y a 4 mois
..
composable_kernel @ 8182976c37 d8f104e97a Support AMD ROCm on FlashAttention 2 (#1010) il y a 4 mois
cutlass @ 756c351b49 74b0761ff7 [FA3] BF16 forward il y a 5 mois
flash_attn 65f723bb9a Split bwd into more .cu files to speed up compilation il y a 4 mois
flash_attn_ck d8f104e97a Support AMD ROCm on FlashAttention 2 (#1010) il y a 4 mois
ft_attention 50896ec574 Make nvcc threads configurable via environment variable (#885) il y a 9 mois
fused_dense_lib 50896ec574 Make nvcc threads configurable via environment variable (#885) il y a 9 mois
fused_softmax 50896ec574 Make nvcc threads configurable via environment variable (#885) il y a 9 mois
layer_norm 50896ec574 Make nvcc threads configurable via environment variable (#885) il y a 9 mois
rotary 50896ec574 Make nvcc threads configurable via environment variable (#885) il y a 9 mois
xentropy 50896ec574 Make nvcc threads configurable via environment variable (#885) il y a 9 mois