So, I have successfully trained a unet model to segment pictures, the predicted images look reasonable, thanks for Jeremy & everybody else!
My question is: I know that the segmentation info is in the predicted object, but is there an easy way to show what color on the image is what class?