I note that the drag_gan function at draggan.py have a parameter "mask", and it did not use. And ` # if mask is not None: # loss += ((F - F0) * (1 - mask)).abs().mean() * lam` has been mask“#”
I note that the drag_gan function at draggan.py have a parameter "mask", and it did not use.
And
# if mask is not None: # loss += ((F - F0) * (1 - mask)).abs().mean() * lamhas been mask“#”