Tri Dao
|
6807b1ea37
Longest-processing-time-first scheduler for causal
|
1 gün önce |
Tri Dao
|
df96486c31
Decode: varlen, paged KV, leftpad
|
1 ay önce |
Tri Dao
|
6e8b25e426
Refactor
|
2 ay önce |
Ying Zhang
|
dff976a84a
fixes
|
3 ay önce |
jayhshah
|
5018ac6ac5
Fp8 kernel with "in-kernel" transpose of V in producer (#1100)
|
4 ay önce |
Tri Dao
|
74b0761ff7
[FA3] BF16 forward
|
5 ay önce |
Tri Dao
|
7f67966cc7
FA3 initial code release
|
5 ay önce |