pytorch-semantic-segmentation
pytorch-semantic-segmentation copied to clipboard
Assertion errors
trafficstars
I am facing size mismatch issues while training cityscapes. Primarily, the labels are of shape (16L, 512L, 256L, 3L) while the outputs are of shape (16L, 19L, 512L, 256L).
Also, how was this resolved?
The label batch should be in shape (16, 512, 256) and the output batch should be in shape (16, 19, 512, 256). Check whether you loaded the wrong label image.
For this issue you should check whether torchvision.transforms.ToTensor is used for the label image.