You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This is the `readme' for my python implementation of binary classification, using a Gaussian Process with Laplace's approximation.
Included is a demonstration taken from `Rasmussen, Williams 2006 Gaussian Processes For Machine Learning', see `http://www.gaussianprocess.org/gpml/'.
The demo uses machine learning to classify images of hand-written digits collected by United States Postal Service, in particular images of `3' and `5's.
The USPS data is taken from `http://www.gaussianprocess.org/gpml/data/'.
setup.sh
--------
Run this script first to download and de-compress the demo data.
demo.sh
--------
Run this script in the shell to view the demo.
The model is trained and then tested interactively.
Training should take less than two minutes to complete on a modern laptop.