RuntimeError: Sizes of tensors must match except in dimension 2. Got 74 and 75 (The offending index is 0)

[balalolaq]

运行时报这种错，是需要训练集中各个图像的size必须一致吗？

[stellating]

可能是由于输入的图像尺寸在不同的卷积操作之后变得尺寸不一样，我原来的图像是700700就报错，但把输入的图像统一resize成256256之后就可以了

[FilippoHei]

I get a similar error (see below), when I try to run it on images of the input size 1080x3840.
Does this Unet only accept quadratic input shapes or how can I avoid this error? Thanks in advance

RuntimeError: Sizes of tensors must match except in dimension 2. Got 134 and 135 (The offending index is 0)

I also rescaled my images and masks to a multiple of 16, but got the same error...