david/flash-attention

Mirror von https://github.com/Dao-AILab/flash-attention

Autor	SHA1 Nachricht	Datum
Tri Dao	0e8c46ae08 Run isort and black on test files	vor 1 Jahr
Tri Dao	88173a1aaf [FusedDense] Support relu, rename FusedDenseGeluDense -> FusedMLP	vor 1 Jahr
Tri Dao	93383bd55b [TP] Implement TensorParallel without sequence parallel	vor 1 Jahr
Tri Dao	ef1ba918c6 [GPT] Refactor function to shard state_dict for TensorParallel	vor 1 Jahr
Tri Dao	b4018a5028 Implement Tensor Parallel for GPT model	vor 2 Jahren