Neil Tenenholtz 7153673c1a Fix swiglu backwards return type (#1337)		2024-11-15 16:23:40 -08:00
..
layers	[Rotary] Support qkv block layout from GQA	2024-09-11 10:39:58 -07:00
losses	[CrossEntropy] Support precomputed LSE	2024-09-08 09:24:43 -07:00
models	minify torch.torch.int32 to torch.int32 (#1237)	2024-09-18 00:32:59 -07:00
modules	Fix: check the type of max_seqlen_k instead of checking max_seqlen twice (#1127)	2024-08-05 08:59:23 -07:00
ops	Fix swiglu backwards return type (#1337)	2024-11-15 16:23:40 -08:00
utils	Update citation	2024-05-26 16:09:03 -07:00
__init__.py	[CI] Pytorch 2.5.1 does not support python 3.8	2024-11-12 20:02:13 -08:00
bert_padding.py	minify torch.torch.int32 to torch.int32 (#1237)	2024-09-18 00:32:59 -07:00
flash_attn_interface.py	Add custom ops for compatibility with PT Compile (#1139)	2024-09-17 19:49:26 -07:00
flash_attn_triton_og.py	Run isort and black on python files	2023-08-18 14:22:11 -07:00
flash_attn_triton.py	Run isort and black on python files	2023-08-18 14:22:11 -07:00
flash_blocksparse_attention.py	minor changes to unpad_input test util func	2024-09-16 14:24:11 -07:00
flash_blocksparse_attn_interface.py	Run isort and black on python files	2023-08-18 14:22:11 -07:00
fused_softmax.py	Run isort and black on python files	2023-08-18 14:22:11 -07:00
pyproject.toml	Move pyproject.toml to flash-attn and tests dir to avoid PEP 517	2023-08-25 15:05:28 -07:00