Skip to content
Methods to help with logging GPflow optimisation.
Branch: master
Clone or download
rhaps0dy and markvdw Bugfix: global_step op created at each step with ScipyOptimizer (#8)
ManagedOptimisation used to create op to increase global_step at every iteration
when used with ScipyOptimizer
Latest commit 646def9 May 4, 2018
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
gpflow_monitor
notebooks Updated example notebook for gpflow changes. Nov 28, 2017
.gitignore Code clean-up. Oct 25, 2017
LICENSE Create LICENSE Oct 25, 2017
README.md Initial commit. Oct 25, 2017
setup.py Initial commit. Oct 25, 2017

README.md

gpflow-monitor

Tools for logging an optimisation procedure, specifically designed to work well with GPflow models.

A very similar structure should work well for general optimisation, which is what opt_tools attempts. This repo is the continuation specifically for GPflow models.

Setup

python setup.py develop

Usage

This code works with the new branch of GPflow. See ./notebooks/ for usage examples.

Support

This code is actively used by researchers who heavily use GPflow. We will endeavour to

  • make sure the code is well-tested,
  • the tests will break whenever a breaking change is made to GPflow.

However we cannot put the same effort into maintaining it as is done for GPflow, simply because logging is not always the priority for our daily activities. If something is broken, or if you want to improve something, we will be happy to consider pull requests.

You can’t perform that action at this time.