Table 6 Mean absolute error on MPF-2021.2.8 subset

From: Materials Graph Library (MatGL), an open-source graph deep learning library for materials science and chemistry

Model

Energy (meV atom−1)

Force (eV Å−1)

Stress (GPa)

M3GNet

19.817/22.558/23.037

0.063/0.072/0.071

0.259/0.399/0.351

TensorNet

28.628/29.708/30.313

0.078/0.083/0.083

0.361/0.471/0.394

CHGNet

17.256/18.226/19.897

0.054/0.061/0.061

0.254/0.347/0319

  1. The numbers are the calculated energy, force and stress mean absolute errors (MAEs) of M3GNet, TensorNet, and CHGNet compared to DFT. The numbers are listed in the order of training, validation, and test. The dataset was divided into training, validation, and test sets with a split ratio of 0.9, 0.05, and 0.05, respectively.
Back to article page