Model
M3GNet
A general-purpose machine learning potential function model based on graph neural networks, used for interatomic potential prediction and structure optimization in materials science.
Timeline (last 90 days)
Oct 27
Machine learning potentials (MLP) are a class of models trained on first-principles data to approximate potential energy surfaces, with typical representatives including MACE, CHGNet, M3GNet, and the ORB series.
Unverified50%