Website | Documentation | Manual | Glossary | Tips & Tricks
GPflow is a package for building Gaussian process models in python, using TensorFlow. It was originally created and is now managed by James Hensman and Alexander G. de G. Matthews. The full list of contributors (in alphabetical order) is Alexander G. de G. Matthews, Alexis Boukouvalas, Artem Artemev, Daniel Marthaler, David J. Harris, Hugh Salimbeni, Ivo Couckuyt, James Hensman, Keisuke Fujii, Mark van der Wilk, Mikhail Beck, Pablo Leon-Villagra, Rasmus Bonnevie, ST John, Tom Nickson, Valentine Svensson, Vincent Dutordoir, Zoubin Ghahramani. GPflow is an open source project so if you feel you have some relevant skills and are interested in contributing then please do contact us.
The development of the GPflow 2.0 moved to the develop branch. The old GPflow 1.0 branch has been transferred to the develop-1.0. We stop supporting GPflow 1.0 in terms of new feature and we will accept only bug fixes into develop-1.0 branch.
GPflow implements modern Gaussian process inference for composable kernels and likelihoods. The online documentation (develop)/(master) contains more details. The interface follows on from GPy, and the docs have further discussion of the comparison.
GPflow 2.0 uses TensorFlow 2.0 for running computations, which allows fast execution on GPUs, and uses Python 3.6.
With the release of tensorflow 2.0 and tensorflow-probability 0.8, you should
only need to run
pip install -e .in a check-out of the develop-2.0 branch of the GPflow github repository.
There is an "Intro to GPflow 2" Jupyter notebook in doc/source/notebooks/intro_to_gpflow2.ipynb". Not all tutorial notebooks may be running on gpflow2` already, but we are working on it!
Please use GitHub issues to start discussion on the use of GPflow. Tagging enquiries discussion helps us distinguish them from bugs.
All constructive input is gratefully received. For more information, see the notes for contributors.
GPflow heavily depends on TensorFlow and as far as TensorFlow supports forward compatibility, GPflow should as well. The version of GPflow can give you a hint about backward compatibility. If the major version has changed then you need to check the release notes to find out how the API has been changed.
Unfortunately, there is no such thing as backward compatibility for GPflow models, which means that a model implementation can change without changing interfaces. In other words, the TensorFlow graph can be different for the same models from different versions of GPflow.
A few projects building on GPflow and demonstrating its usage are listed below.
| Project | Description |
|---|---|
| GPflowOpt | Bayesian Optimization using GPflow. |
| VFF | Variational Fourier Features for Gaussian Processes. |
| Doubly-Stochastic-DGP | Deep Gaussian Processes with Doubly Stochastic Variational Inference. |
| BranchedGP | Gaussian processes with branching kernels. |
| heterogp | Heteroscedastic noise for sparse variational GP. |
| widedeepnetworks | Measuring the relationship between random wide deep neural networks and GPs. |
| orth_decoupled_var_gps | Variationally sparse GPs with orthogonally decoupled bases |
Let us know if you would like your project listed here.
To cite GPflow, please reference the JMLR paper. Sample Bibtex is given below:
@ARTICLE{GPflow2017,
author = {Matthews, Alexander G. de G. and {van der Wilk}, Mark and Nickson, Tom and
Fujii, Keisuke. and {Boukouvalas}, Alexis and {Le{\'o}n-Villagr{\'a}}, Pablo and
Ghahramani, Zoubin and Hensman, James},
title = "{{GP}flow: A {G}aussian process library using {T}ensor{F}low}",
journal = {Journal of Machine Learning Research},
year = {2017},
month = {apr},
volume = {18},
number = {40},
pages = {1-6},
url = {http://jmlr.org/papers/v18/16-537.html}
}