graph-pes is a framework built to accelerate the development of machine-learned potential energy surface (PES) models that act on graph representations of atomic structures.
Features
Interatomic Potentials - Graph-based machine learning models for potential energy surfaces.