Kernels
attention
flash-attention
flash-attn-4
sm120
sm121
blackwell
rtx5090
rtx-pro-6000
dgx-spark
cute-dsl
Instructions to use SecondNatureComputing/flash-attn-4-sm120 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Kernels
How to use SecondNatureComputing/flash-attn-4-sm120 with Kernels:
# !pip install kernels from kernels import get_kernel kernel = get_kernel("SecondNatureComputing/flash-attn-4-sm120") - Notebooks
- Google Colab
- Kaggle
Update repo-id in build.toml after migration to SNC org
Browse files- build.toml +1 -1
build.toml
CHANGED
|
@@ -6,7 +6,7 @@ license = "BSD-3-Clause"
|
|
| 6 |
version = 0
|
| 7 |
|
| 8 |
[general.hub]
|
| 9 |
-
repo-id = "
|
| 10 |
|
| 11 |
[general.cuda]
|
| 12 |
# flash-attn-4 requires CUDA 12.8+ per upstream.
|
|
|
|
| 6 |
version = 0
|
| 7 |
|
| 8 |
[general.hub]
|
| 9 |
+
repo-id = "SecondNatureComputing/flash-attn-4-sm120"
|
| 10 |
|
| 11 |
[general.cuda]
|
| 12 |
# flash-attn-4 requires CUDA 12.8+ per upstream.
|