zijundeng/pytorch-semantic-segmentation

consistent with input size when it is odd during test?

micklexqg opened this issue · 0 comments

I wonder that if the input size of image is odd, is the output size consistent with it during test?