Table 1 Benchmarking model performance against existing models.

From: Formation energy prediction of crystalline compounds using deep convolutional network learning on voxel image representation

 

Shallow CNN (this work)

Deep CNN (this work)

ALIGNN

CGCNN

ElemNet

Roost

MAE (eV per atom)

0.337

0.046

0.022

0.039

0.1

0.06

  1. Comparison of the formation energy MAE of different models (ElemNet16, Roost17, ALIGNN23, and CGCNN18) and this work.
Back to article page