fa2.py contains 3 different attention implementations:
- Vanilla attention in PyTorch
- FlashAttention2 in PyTorch
- FlashAttention2 in Triton
-
Notifications
You must be signed in to change notification settings - Fork 0
seanbae/fa2
Folders and files
| Name | Name | Last commit message | Last commit date | |
|---|---|---|---|---|
Repository files navigation
About
FlashAttention2 in Triton and PyTorch
Resources
Stars
Watchers
Forks
Releases
No releases published