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