search

odegeasslbc / FastGAN-pytorch

Official implementation of the paper "Towards Faster and Stabilized GAN Training for High-fidelity Few-shot Image Synthesis" in ICLR 2021
GNU General Public License v3.0
600 stars 100 forks source link

Question w.r.t SEblock. #40

Closed SHShim0513 closed 2 years ago

SHShim0513 commented 2 years ago

I have a question w.r.t SEblcok.

Do you have any specific reason why bias option is set to "False" in conv layer of SEblock?

I know that the FastGAN and MobileNet also have the same option.

Thank you.

odegeasslbc commented 2 years ago

Hi,

I generally find that when training GANs, setting conv layer's bias to False lead to better convergence. I think it was a discovery since early DCGAN era, like in 2017, but I cannot recall exactly, you can also check the official DCGAN repo.

SHShim0513 commented 2 years ago

Thank you for your kind answer!

I'll check it.