Closed yucornetto closed 4 years ago
Thanks! Here's the laplacian loss, https://gist.github.com/MarcoForte/a07c40a2b721739bb5c5987671aa5270 I based it off of this https://gist.github.com/alper111/b9c6d80e2dba1ee0bfac15eb7dad09c8 but I modified the code to support masking.
Thanks! That's very helpful
I also want to add that the original code only worked for square images but my version works for rectangular ones also.
@MarcoForte What do you mean by 'masking' in your previous comment?
what is weight in the laplacian loss?
@yucornetto and @bluesky314 were you able to figure out what is meant by masking or weight in the laplacian loss ?
Thanks for the great work! Your results are really impressive. I noticed that you suggested using laplacian loss for training. May I ask for the implementation details of it?