Closed YuMeng2v closed 2 years ago
In colorization-pytorch, the siggraph generator model forward function is written without [normalization]. But the demo code seems to be the other way, it normalize the data. Anyone who has test which is the real training way?
In colorization-pytorch, the siggraph generator model forward function is written without [normalization]. But the demo code seems to be the other way, it normalize the data. Anyone who has test which is the real training way?