Runtime error: The size of tensor a (275) must match the size of tensor b (276) at non-singleton dimension 3

[cherrysherryplus]

When one of the input image's resolution (width or height) is not divisiable by 8 (or the downsampling rate in the AER-Net forward process), the error happens.

---

The error message

The input clear image (550 x 975)

Should I do some resize or crop work with my own dataset both when training and testing? What should I do to make images with random size as a legal input?

Thanks a lot.

[cherrysherryplus]

# The difference between x_down2 and x_up1 causes the error
input haze size: (550, 975)
x_down2 haze size: torch.Size([1, 128, 488, 275])
x_up1 haze size: torch.Size([1, 128, 488, 276])