yangcf10

Hi, first thanks for the great work. I do have a question about the edge loss used in the training (defined in networks/losses.py). I am wondering how does it work...