FTransGAN in PyTorch

This is the implementation for the paper:

Few-shot Font Style Transfer between Different Languages

in Proc. of Winter Applications on Computer Vision WACV’21, Jan., 2021. Paper.

Requirements

• Linux
• CPU or NVIDIA GPU + CUDA CuDNN
• Python 3
• torch>=0.4.1
• torchvision>=0.2.1
• dominate>=2.3.1
• visdom>=0.1.8.3

How to use

• Download the dataset.

• Unzip it to ./datasets/

• To view training results and loss plots, run python -m visdom.server and click the URL http://localhost:8097.

• Train the model

bash ./train.sh

• Test

bash ./test.sh

• Evaluate

bash ./evaluate.sh

• scripts.sh integrate train.sh, test.sh, and evaluate.sh

bash ./scripts.sh

Gallery

Acknowledgements

Code derived and reshaped from:

pix2pix