https://github.com/GPflow/GPflow
Revision 5b16a12e3a21e1435badfe99c4821444260df2fa authored by John Bradshaw on 23 October 2017, 10:32:06 UTC, committed by John Bradshaw on 23 October 2017, 10:32:06 UTC
the future. * EP while loop broken out and made into seperate class to allow it to be easy to reuse elsewhere. Tidy up of the variable names and made more general. Tests written to check that it can recover the true solution when the approximating factors have the same solution as the true solution. * Central moment functions moved out of EP class and live elsewhere. Step function, Gaussian and Gaussian CDF moments implemented all with tests. * Likelihoods now are associated with a method that returns their central moments generator function. Only implemented for Bernoulli but this allows one to store moment logic with the likelihood function which is more where it belongs. * Updated the VGP versus EP binary classification example script to use this new source structure. This has remained in the docs so have not written tests for this part.
1 parent f46eeb3
Tip revision: 5b16a12e3a21e1435badfe99c4821444260df2fa authored by John Bradshaw on 23 October 2017, 10:32:06 UTC
Refactor of EP implementation to modularise and make more extensible in
Refactor of EP implementation to modularise and make more extensible in
Tip revision: 5b16a12
README.md
# GPflow
GPflow is a package for building Gaussian process models in python, using [TensorFlow](http://www.tensorflow.org). It was originally created and is now managed by [James Hensman](http://www.lancaster.ac.uk/staff/hensmanj/) and [Alexander G. de G. Matthews](http://mlg.eng.cam.ac.uk/?portfolio=alex-matthews).
The full list of [contributors](http://github.com/GPflow/GPflow/graphs/contributors) (in alphabetical order) is Artem Artemev, Rasmus Bonnevie, Alexis Boukouvalas, Ivo Couckuyt, Keisuke Fujii, Zoubin Ghahramani, David J. Harris, James Hensman, Pablo Leon-Villagra, Daniel Marthaler, Alexander G. de G. Matthews, Tom Nickson, Valentine Svensson, Mark van der Wilk. 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.
[](https://codeship.com/projects/147609)
[](https://travis-ci.org/GPflow/GPflow)
[](http://codecov.io/github/GPflow/GPflow?branch=master)
[](http://gpflow.readthedocs.io/en/latest/?badge=latest)
# What does GPflow do?
GPflow implements modern Gaussian process inference for composable kernels and likelihoods. The [online user manual](http://gpflow.readthedocs.io/en/latest/) contains more details. The interface follows on from [GPy](http://github.com/sheffieldml/gpy), for more discussion of the comparison see [this page](http://gpflow.readthedocs.io/en/latest/intro.html#what-s-the-difference-between-gpy-and-gpflow).
# Install
## 1) Quick install
GPflow can be installed by cloning the repository and running
```
pip install .
```
in the root folder. This also installs required dependencies including TensorFlow. When GPU support is needed, a manual installation of TensorFlow is recommended (next section), as one cannot rely on pip to get this running.
## 2) Alternative method
A different option to install GPflow requires installation of TensorFlow first. Please see instructions on the main TensorFlow [webpage](https://www.tensorflow.org/versions/r1.0/get_started/get_started). You will need at least version 1.0 (we aim to support the latest version). We find that for most users pip installation is the fastest way to get going. Then, for those interested in modifying the source of GPflow, we recommend
```
python setup.py develop
```
but installation should work well too:
```
python setup.py install
```
You can run the tests with `python setup.py test`.
Version history is documented [here.](https://github.com/GPflow/GPflow/blob/master/RELEASE.md)
# Deprecation
Python package name `GPflow` is no longer supported, it has been changed to lower-case name `gpflow`. You can adapt your code to new renamed package by running these commands:
```bash
## All files will be backed-up with `.bak` suffix
sed -i '.bak_import' 's/^\(import *\) GPflow/\1 gpflow/g' ./project-path
sed -i '.bak_from' 's/^\(from *\) GPflow/\1 gpflow/g' ./project-path
sed -i '.bak_dot' 's/GPflow\(\.[a-zA-Z0-9]\)/gpflow\1/g' ./project-path
```
## Docker image
We also provide a [Docker image](https://hub.docker.com/r/gpflow/gpflow/) which can be run using
```
docker run -it -p 8888:8888 gpflow/gpflow
```
Code to generate the image can be found [here](Dockerfile)
# Getting help
Please use gihub issues to start discussion on the use of GPflow. Tagging enquiries `discussion` helps us distinguish them from bugs.
# Contributing
All constuctive input is gratefully received. For more information, see the [notes for contributors](contributing.md).
# Projects using GPflow
A few projects building on GPflow and demonstrating its usage are listed below.
| Project | Description |
| --- | --- |
| [GPflowOpt](https://github.com/GPflow/GPflowOpt) | Bayesian Optimization using GPflow. |
| [VFF](https://github.com/jameshensman/VFF) | Variational Fourier Features for Gaussian Processes. |
| [Doubly-Stochastic-DGP](https://github.com/ICL-SML/Doubly-Stochastic-DGP)| Deep Gaussian Processes with Doubly Stochastic Variational Inference.|
| [BranchedGP](https://github.com/ManchesterBioinference/BranchedGP) | Gaussian processes with branching kernels.|
Let us know if you would like your project listed here.
# Citing GPflow
To cite GPflow, please reference the [JMLR paper](http://www.jmlr.org/papers/volume18/16-537/16-537.pdf). 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}
}
```
Computing file changes ...
