Matmul() using PyTorch's MPS back end is faster than Apple's MLXkevinmartinjose.com 1 pointlonesword1 year ago1 commentSaveHideCopy link On HNComments−drewbitt1yI have noticed the same thing with PyTorch and MLX.
Comments
I have noticed the same thing with PyTorch and MLX.