my_softmax_function / build /torch-cuda /attention_v3.py

Commit History

add built softmax attention kernel
490d9fe

tiagomonteiro0715 commited on