Sparse Gaussian Process Variational Autoencoders

This repository contains the Python implementation of the SGP-VAE, introduced in our paper.

The main components of the repository are:

• sgpvae: the implementation of the SGP-VAE and partial inference networks.
• experiments: code for running the experiments detailed in the paper.
• data: code for installing the datasets used in the experiments.

Installation and Dependencies

This code is implemented in Python 3.8. To install, simply

pip install sgpvae

Contact

Please do feel free to use/extend this code for your own research. If you have any questions, or would like to report any issues, please open an issue on the issues tracker or contact me at mca39@cam.ac.uk.