I\'m trying to implement an U-NET for pupil segmentation by using the open-source code of DEEPVOG. the idea is to train the model with my personal dataset... I have 500 gray ima
