david/flash-attention: flash-attention from https://github.com/Dao-AILab/flash-attention @ changes_for_fp8

miroir de https://github.com/Dao-AILab/flash-attention

Markus Krimmel 6bbc532388 fix: cast the alibi slopes to torch.float32 (#846)		il y a 9 mois
..
__init__.py	ece539abd6 Add __init__.py files to subdirectories for installation	il y a 2 ans
block.py	abbc131173 [LayerNorm] Switch from CUDA to Triton implementation	il y a 1 an
embedding.py	f1a73d0740 Run isort and black on python files	il y a 1 an
mha.py	6bbc532388 fix: cast the alibi slopes to torch.float32 (#846)	il y a 9 mois
mlp.py	c3b2196652 Add Alibi to MHA, test with Baichuan-13B	il y a 1 an