david/flash-attention

огледало од https://github.com/Dao-AILab/flash-attention

Аутор	SHA1 Порука	Датум
Tri Dao	5400fdc4ac [CE] Implement CrossEntropyLoss in Triton	пре 1 година
Tri Dao	43ab0b5205 Mention that some CUDA extensions have only been tested on A100s	пре 2 година
Tri Dao	fa6d1ce44f Add fused_dense and dropout_add_layernorm CUDA extensions	пре 2 година