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