david/flash-attention

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

Autor	SHA1 Mensaxe	Data
Tri Dao	0e8c46ae08 Run isort and black on test files	hai 1 ano
Tri Dao	88173a1aaf [FusedDense] Support relu, rename FusedDenseGeluDense -> FusedMLP	%!s(int64=2) %!d(string=hai) anos
Tri Dao	e68ebbe89a Simplify FusedDense	%!s(int64=2) %!d(string=hai) anos
Tri Dao	fa6d1ce44f Add fused_dense and dropout_add_layernorm CUDA extensions	%!s(int64=2) %!d(string=hai) anos