An implementation of Stochastic Gradient Hamiltonian Monte Carlo by Gilad Amitai and Beau Coker.
The folder logistic_regression contains all the code for producing figures. In particular, the Ipython notebook logistic_regression_optimized.ipynb contains the optimized C++ version using Pybind11.
The report can be reproduced by running the Makefile in the Report folder.
The package sghmc can be installed by cloning the repository and running this line in the main folder: python setup.py install.