How is your network initialized?

yuanming-hu / exposure

Learning infinite-resolution image processing with GAN and RL from unpaired image datasets, using a differentiable photo editing model.

MIT License

767 stars 156 forks source link

Open princexus opened 5 years ago

princexus commented 5 years ago

Hello, thanks for your opening codes. I want to know how is your network initialized before training? Thanks Looking forward to your reply

yuanming-hu commented 5 years ago

I think that's just a default Xavier initialization in tensorflow.

princexus commented 5 years ago

Thanks for your reply. Is it possible if I want to fine-tune the network based on your well-trained model?If possible,which codes should I modify?