purpose of train_resnet

pbaylies / stylegan-encoder

StyleGAN Encoder - converts real images to latent space

Other

740 stars 180 forks source link

purpose of train_resnet #28

Closed segalinc closed 5 years ago

segalinc commented 5 years ago

Hi and thanks for your contribution. What is the purpose of training or using resent/affnet model in encode_images if you already have the vgg perceptual model?

MartiGrau commented 5 years ago

Hi @cristinasegalin , you could check this video https://www.youtube.com/watch?v=dCKbRCUyop8 from 15:00 in advance in order to understand the ResNet or EfficientNet use.

pbaylies commented 5 years ago

Seconded, that's a great tutorial video! The resnet just serves to provide a good initial guess for the encoder.

segalinc commented 5 years ago

Nice, I guessed that was the reason, wanted to confirm. thanks. I am curious to see how fast the optimization turns using this guess compared to now using it.