Skip to content

Latest commit

 

History

History
11 lines (7 loc) · 835 Bytes

README.md

File metadata and controls

11 lines (7 loc) · 835 Bytes

HamNet

ATTENTION: This project will be no more updated, but it can still be used for reproduction. If you want to reuse our code, please visit MoleculeClub.

Conformation-Guided Molecular Representation with Hamiltonian Neural Networks. In ICLR 2021

Note that QM9 dataset is not uploaded to this project due to 100M capacity limit, but it's available on MoleculeNet Dataset. After downloading QM9(structure), just unzip it and copy gdb9.sdf to data/gdb9/ and everything is ok.

You can run the xxx.slurm to reproduce the results claimed in paper.

  • Run fit.slurm first to train a HamEngine (run fit2.slurm if you don't need adj3_loss)
  • Run molecule property prediction jobs with the trained HamEngine (molecule conformation generator)