Training on new dataset

[XinyangHan]

Hi, I really appreciate the work you have done and I want to train your color module on new dataset. In CPM/Color/Readme.md, it says to re-train model on new dataset, one should follow the the instruction on BeautyGAN. I am wondering whether this means I should use the BeautyGAN network to train on new dataset and use the checkpoints from this as the new color module checkpoints, or I can simply train the color module in this CPM network?

[thaoshibe]

Hi, thanks for your interest! The CPM's Color network is basically the same as BeautyGAN network. The only difference is the dataset. While BeautyGAN is trained on original images, CPM's Color is trained on UV-map of original images.

So, if you want to train Color Module on a new dataset, please use the BeautyGAN network to train.

Hope this helps. Feel free to ask if you have any further questions.

[XinyangHan]

Thanks for your answering!

[XinyangHan]

Problem solved!