How to reproduce the paper code by using pytorch?

Is there anyone interested in the paper “Optimizing Through Learned Errors for Accurate Sports Field Registration”. Would you like to discuss how to reproduce the paper code by using pytorch?