Introduction

What is MACE ?

MACE provides fast and accurate machine learning interatomic potentials with higher order equivariant message passing. For faster training and inference using the cuEquivariance library, please read the CUDA acceleration with cuequivariance library section.

Also available:

• MACE in JAX, currently about 2x times faster at evaluation, but training is recommended in PyTorch for optimal performances and features.

• MACE layers, for constructing higher order equivariant graph neural networks for arbitrary 3D point clouds.