vladgaus / twingan

Twin-GAN -- Unpaired Cross-Domain Image Translation with Weight-Sharing GANs

TwinGAN -- Unsupervised Image Translation for Human Portraits

Extra materials:

Presentation Slides at Anime Expo 2018

Technical report coming soon!

Use Pretrained Model.

For a pretrained model, you can download it here:

Run the following command to translate the demo inputs.

python inference/image_translation_infer.py \
--model_path="/PATH/TO/MODEL/256/"
--image_hw=256
--input_tensor_name="sources_ph"
--output_tensor_name="custom_generated_t_style_source:0"
--input_image_path="./demo/inference_input/"
--output_image_path="./demo/inference_output/"

The input_image_path can be either one single image or a path containing images.

Training

Download CelebA and the Getchu dataset by following the datasets guide. Then train your model using script from the training guide.

Technical report.

Please refer to the blog at the moment. A more technical report will be shared soon!

Related works

Our idea of using adaptive normalization parameters for image translation is not unique. To the best of our knowledge, at least two more work have similar ideas: MUNIT and EG-UNIT. Our model is developed around the same time period with these models.

Some key differences between our model and the two mentioned are -- we find UNet to be extremely helpful in maintaining semantic correspondence across domain, and we found that tying all convolution filter weights speeds up training while maintaining the same output quality.

Reference

A lot of the code are adapted from online. Here is a non-exhaustive list of the repos where I borrowed code from extensively.

TF-Slim image models library

PGGAN

Disclaimer

This personal project is developed and open sourced when I am working for Google, therefore you see Copyright 2018 Google LLC in each file. This is not an officially supported Google product. See License and Contributing for more details.