pytorch implementation for CartoonGAN and CycleGAN for avatar project

the CartoonGAN follows this project
the CycleGAN uses the vanilla cycleGAN project with some minor tweaks

Folder structure

The following shows data folder structure.

── data
   ├── src_data # src data (not included in this repo)
   │   ├── train 
   │   └── test
   └── tgt_data # tgt data (not included in this repo)
       ├── train 
       └── pair # edge-promoting results to be saved here, only use for CartoonGAN

The training data needs to be downloaded from this link and place inside the data

Usage for CartoonGAN

• Download VGG19 VGG19 put the model under pre_trained_VGG19_model_path folder
• Run CartoonGAN.ipynb

Usage for CycleGAN

• Run CycleGAN.ipynb

Usage for Pix2Pix (Supervised GAN)

• Use the photo-manga pair images, the dataset can be downloaded from here
• Run Pix2Pix.ipynb

Development Environment

• NVIDIA Tesla P100
• cuda 8.0
• python 3.5.5
• pytorch 0.4.1
• torchvision 0.2.1
• opencv-python 3.4.3.18