code based on junyanz/pytorch-CycleGAN-and-pix2pix
- efficient-det as discriminator backbone
- trivial upsample generator
pip install -r requirements.txt
To view training results and loss plots, run python -m visdom.server and click the URL http://localhost:8097.
If you run without it is fine. The code will run the command automaitc
python train.py --dataroot ./dataset/selfie2anime --name selfie2anime_ustyle --model ustyle
python train.py --dataroot ./dataset/selfie2anime --name selfie2anime_ustyle --model ugatit --light
python train.py --dataroot ./dataset/selfie2anime --name selfie2anime_ustyle --model ustyle --continue_train
python train.py --data_root ./horse2zebra --name h2z_cyclegan --model cycle_gan
python test.py --dataroot ./dataset/selfie2anime --name selfie2anime_ustyle --model ustyle
python test.py --data_root ./horse2zebra --name h2z_cyclegan --model cycle_gan