Codes

RxnRep

RxnRep is a training method to leverage self-supervised contrastive pretraining to learn chemical reaction representations. It enables the building of graph neural networks for small chemical reaction data.

BonDNet

BonDNet is a graph neural network model for the prediction of bond dissociation energies (BDEs). It can be applied to both homolytic and heterolytic bond dissociations for molecules of any charge.

KLIFF

A KIM-based Learning-integrated Fitting Framework to train both physics-motivated and machine learning interatomic potentials.

pair dirp

LAMMPS implementation of the Dihedral-angle-corrected registry-dependent interlayer potential (DRIP) for multilayer graphene structures.

kimpy

A Python interface to the KIM API for atomistic simulations.