Variational Autoencoder with Arbitrary Conditioning for Image Inpainting in TensorFlow 2.0

An implementation of the paper "Variational Autoencoders with Arbitrary Conditional" in TensorFlow 2.0. This implementation is based on the original code by the paper authors (link)

Installation

pip install virtualenv
virtualenv -p python3 .env
source .env/bin/activate
pip install -r requirements.txt

How to run

Prepare CelebA dataset

python run.py --mode prepare

Debug

python run.py --mode debug

Training

python run.py --mode train

Inpainting

python run.py --mode inpaint

Some inpainted results on CelebA

We train the model in 40 epochs on the CelebA dataset. Below are some inpainted results generated by our trained model. The first column indicates input images with arbitrary masks to be inpainted. The last column indicates ground truth images. And the remaining columns are inpainted images.

References

1. Oleg Ivanov, Michael Figurnov, Dmitry Vetrov. Variational Autoencoder with Arbitrary Conditioning. ICLR 2019 (arxiv)