Published May 20, 2022 | Version v4
Preprint Open

OrbNet-Equi: Informing geometric deep learning with electronic interactions to accelerate quantum chemistry

Creators

  • 1. California Institute of Technology
  • 2. Entos, Inc.
  • 3. California Institute of Technology; NVIDIA
  • 4. Entos, Inc.; California Institute of Technology

Description

Source data for the manuscript "Informing geometric deep learning with electronic interactions to accelerate quantum chemistry"; The SDC21 training dataset; Neural network code and evaluation examples.

 

Version 2 note: The license is updated. Typos in README.md and minor numerical errors in the U_0 predictions have been corrected.

Version 3/4 note: Additional scripts and summary dataframes for cube-based charge density analysis.

Files

all.zip

Files (19.1 GB)

Name Size Download all
md5:8fb2be187729ac0b94923c13bf095a7f
19.1 GB Preview Download