Repo to store interatomic potentials developed through machine learning and corresponding training data
An AMP neural network model for the lithium-graphite system. The model obtains an error of () meV/atom on the test set. Total dataset contains 9189 structures and corresponding energies/forces.