pytorch-deeplab-resnet
pytorch-deeplab-resnet copied to clipboard
Why change the image channel order after `cv2.imread`
Since cv2.imread
loaded images are already in BGR
format, which is the correct input format for caffe models, why change it here?