thomaspinder / SteinGP

Code to fit Gaussian processes using Stein variational gradient descent. Code is writted in Tensorflow 2.0 and GPFlow 2.

Stein Variational Gaussian Processes

Installation

git clone https://github.com/thomaspinder/SteinGP.git
cd SteinGP/
virtualenv venv --python=python3
source venv/bin/activate
pip install -r requirements.txt

Running experiments

To run experiments it is necessary to install the SteinGP package locally. This can be done by running python setup.py develop from the repository's root directory. To check for a fully working installation, run python -m tests.run_tests.