david/flash-attention

mirrorاز https://github.com/Dao-AILab/flash-attention

نویسنده	SHA1 پیام	تاریخ
Tri Dao	36587c01cb [LayerNorm] Update layer_norm_linear	9 ماه پیش
Tri Dao	bdcae547c7 [LayerNorm] Don't exit early in the backward pass (fix #781)	11 ماه پیش
Tri Dao	c9861a032d [LayerNorm] Initialize mean and rstd tensor using x.device	1 سال پیش
Tri Dao	f5b308e258 [LayerNorm] Rename layernorm.py -> layer_norm.py	1 سال پیش