It depends on how your UNet architecture is constructed. In the example image you've shown, the output of the first few layers of convolutions is cropped to ...
確定! 回上一頁