 Tri Dao
|
29cdfedd80
Use Bulk reduce instead of TMA for dQaccum, split across WGs
|
hace 1 semana |
|
Tri Dao
|
df96486c31
Decode: varlen, paged KV, leftpad
|
hace 1 mes |
|
Tri Dao
|
6e8b25e426
Refactor
|
hace 2 meses |
 Ying Zhang
|
1c9717d699
address comments
|
hace 2 meses |
|
Ying Zhang
|
be6c1b98c4
small fixes
|
hace 2 meses |
|
Ying Zhang
|
dff976a84a
fixes
|
hace 3 meses |
|
Ying Zhang
|
7b4e68e04f
hopper local attention
|
hace 3 meses |
|
Tri Dao
|
3f6ff1c1c5
Remove struct : cute::aligned_struct to avoid error with gcc 12
|
hace 4 meses |
|
Tri Dao
|
bafe253042
[FA3] Bwd
|
hace 4 meses |
|
Tri Dao
|
7f67966cc7
FA3 initial code release
|
hace 5 meses |
