This repositery tries to closely resemble original pix2pix paper with the same architecture and hyper-parameters, but using Swift for TensorFlow. Results seem to be close to what original authors claimed to achieve, this is what a generator outputs after 7 hours of training using single GTX 1080 card.
In order to run the project you have to just call
swift run -c release
You will need a tensorboard installation in order to run the project.
- Trainer will emit a lot of intermediate results in a calling directory, you can easily comment out that part. It is a workaround because for some reason tensorboard crashes when provided with result images, I'm investigating on the problem.
- You need to specify different URL for
dataset
andGPU index
- This implementation only provides a data loader for facades dataset, but the model is capable of learning on any pix2pix compatible dataset.