StyleGAN2_PyTorch
StyleGAN2_PyTorch copied to clipboard
Why does the activation should product sqrt(2)?
https://github.com/tomguluson92/StyleGAN2_PyTorch/blob/37741fbadc59c2fb2613248aa5d74f5b9d66de73/network/stylegan2.py#L95
Hi, This way is same with original TF version, used in Leaky ReLU.
Hi, This way is same with original TF version, used in Leaky ReLU. Why did you comment on the following code? Is the tanh() necessary? https://github.com/tomguluson92/StyleGAN2_PyTorch/blob/37741fbadc59c2fb2613248aa5d74f5b9d66de73/network/stylegan2.py#L913
Hi, This way is same with original TF version, used in Leaky ReLU. Why did you comment on the following code? Is the tanh() necessary? https://github.com/tomguluson92/StyleGAN2_PyTorch/blob/37741fbadc59c2fb2613248aa5d74f5b9d66de73/network/stylegan2.py#L913
I try to restrict the output data fall in [-1, 1], which the original TF version didn't do that.