depth evaluation

[sotoy]

I came across an issue while trying to evaluate the results after training. The dataset I'm using provides gt depth in the form of grayscale images. As I see, you create .npy files with depth maps in the kitti_raw_loader file. How should I evaluate with gt depth images using your code? Also, I see you give some values such as min_value and max_value by hand. What exactly are those?