microsoft

/

wavlm-base-plus-sv

like 27

I get a lot of errors when running this of the variety: Given input with spatial size (1, 1), kernel_size=(3, 512), dilation=(3, 1), padding=(0, 0), calculated shape of the array of sliding blocks as (-5, 1), but its components must be at least one.

Does this model require a certain shape?