I was reading pytorch docs about Conv2DTranspose on .nn.ConvTranspose2d.html#torch.nn.ConvTranspose2d, and I was curious as to what happens if we increase the stride to be larger than 1. I know how transposed convolution works for stride = 1. Are zeros in the output inserted in every 2nd element when we set stride = 2, or something else happens. Also whats the advantage of upsampling the image this way, rather than using upSampling?

Thanks for help in advance.

I was reading pytorch docs about Conv2DTranspose on https://pytorch./docs/stable/generated/torch.nn.ConvTranspose2d.html#torch.nn.ConvTranspose2d, and I was curious as to what happens if we increase the stride to be larger than 1. I know how transposed convolution works for stride = 1. Are zeros in the output inserted in every 2nd element when we set stride = 2, or something else happens. Also whats the advantage of upsampling the image this way, rather than using upSampling?

Thanks for help in advance.

• pytorch
• conv-neural-network

1 Answer 1

Here you can see the differences between upsampling and ConvTranspose2D: Pytorch Forumns discussion.