How Does StyleGAN 2 Work? Modulated Convolution Tutorial & Code

preview_player
Показать описание
How does StyleGAN 2 work? In the first part of a three part series, I go through the theory behind modulated/demodulated convolution; a replacement for adaptive instance normalization in StyleGAN. I include a high level explanation, then an implementation in Tensorflow 2.0.

Music:

Thanks for watching!
Рекомендации по теме
Комментарии
Автор

These videos have helped me a lot in understanding GANs

brownbadass
Автор

Awesome job! I'm looking forward to part 2.

shelbypiechotta
Автор

Nice.Looking for complete StyleGAN 2 Tutorial

thinkforsociety
Автор

Can you please make a video to chart a roadmap(prerequisites) to create such architectures from scratch?

aromax
Автор

I just want to ask why you don't use reshape directly to fuse batch & channel but change dimension formulation to (B, C, H, W) first ?

And normalize kernel group by norm do not compute variance along the group because the mean is not zero

erknee
Автор

Please, make the explanation about StyleGan-3.

nikitabalabin
join shbcf.ru