Tri Dao e81c237e28 [Cute] Fix sink impl ...

Previously we implemented
    e^x
-----------
 sink + Σ e^x

Now we implemented
    e^x
-----------
 e^sink + Σ e^x

2025-08-05 20:42:41 -04:00

..

__init__.py

Add __init__.py files to subdirectories for installation

2022-11-17 16:55:44 -08:00

benchmark.py

[LayerNorm] Implement residual + LayerNorm/RMSNorm in Triton

2023-11-13 02:04:49 -08:00

distributed.py

Run isort and black on python files

2023-08-18 14:22:11 -07:00

generation.py

Update citation

2024-05-26 16:09:03 -07:00

library.py

[LayerNorm] Don't let torch.compile trace inside _layer_norm_bwd

2025-04-29 23:42:49 -04:00

pretrained.py

[GPT] Fix loading weights from HF hub

2023-08-21 22:56:02 -07:00

testing.py

[Cute] Fix sink impl

2025-08-05 20:42:41 -04:00

torch.py

fix deprecation warning for newer torch versions (#1565)

2025-03-31 18:44:32 -07:00