Skip to content

Top New Best Ask Show Jobs

Matmul() using PyTorch's MPS back end is faster than Apple's MLX | Better HN

Matmul() using PyTorch's MPS back end is faster than Apple's MLX (opens in new tab)

(kevinmartinjose.com)

1 pointslonesword11mo ago1 comments

1 comments

drewbitt11mo ago

I have noticed the same thing with PyTorch and MLX.

j / k navigate · click thread line to collapse