search

f-dangel / unfoldNd

(N=1,2,3)-dimensional unfold (im2col) and fold (col2im) in PyTorch
MIT License
82 stars 6 forks source link

Revert "[REF] Use alternative approach where kernel is repeated C_in * N times" #15

Closed f-dangel closed 3 years ago

f-dangel commented 3 years ago

Reverts f-dangel/unfoldNd#14

The expected speed-up in terms of run time by making more copies of the one-hot kernel did not work.