Data from "A Neural Network Water Model Based on the MB-pol Many-Body Potential"

This dataset contains input files, training data and other files related
to the machine learning models developed during the work by Muniz et al.
In this work, we construct machine learning models based on the MB-pol
many-body model. We find that the training set should include cluster
configurations as well as liquid phase configurations in order to
accurately represent both liquid and VLE properties. The results attest
for the ability of machine learning models to accurately represent
many-body potentials and provide an efficient avenue for water
simulations.