Tensor BNN

This package contains code which can be used to create full Bayesian Neural Networks using Hamiltonian Monte Carlo sampling as proposed by Radford Neal in his thesis “Bayesian Learning for Neural Networks” along with some added features. The package is written in python and uses the packages Tensorflow and Tensorflow-Probability as the framework for the implementation.

For instructions on how to setup this package, click here.

If you would like an explanation of how to use the code, click here.