Open ygtxr1997 opened 2 years ago
Your papar mentions that the model can be extended to video face swapping with two extra losses. But it seems that the losses require video training datasets.
Your papar mentions that the model can be extended to video face swapping with two extra losses. But it seems that the losses require video training datasets.