graph-pes

The `graph-pes` framework is designed to accelerate the development of machine-learned potential energy surface models that utilize graph representations of atomic structures. It allows researchers to train, fine-tune, and deploy models for molecular dynamics simulations and energy calculations.