| .. |
|
composable_kernel @ a9b170b541
|
e2182cc21d
Support page kvcache in AMD ROCm (#1198)
|
6 月之前 |
|
cutlass @ 756c351b49
|
74b0761ff7
[FA3] BF16 forward
|
8 月之前 |
|
flash_attn
|
65f723bb9a
Split bwd into more .cu files to speed up compilation
|
8 月之前 |
|
flash_attn_ck
|
e2182cc21d
Support page kvcache in AMD ROCm (#1198)
|
6 月之前 |
|
ft_attention
|
50896ec574
Make nvcc threads configurable via environment variable (#885)
|
1 年之前 |
|
fused_dense_lib
|
50896ec574
Make nvcc threads configurable via environment variable (#885)
|
1 年之前 |
|
fused_softmax
|
50896ec574
Make nvcc threads configurable via environment variable (#885)
|
1 年之前 |
|
layer_norm
|
50896ec574
Make nvcc threads configurable via environment variable (#885)
|
1 年之前 |
|
rotary
|
50896ec574
Make nvcc threads configurable via environment variable (#885)
|
1 年之前 |
|
xentropy
|
50896ec574
Make nvcc threads configurable via environment variable (#885)
|
1 年之前 |
rocking
e2182cc21d
Support page kvcache in AMD ROCm (#1198)