This repository contains the scripts for accessing QO2Mol dataset, the large-scale quantum chemistry dataset with 20 million conformers, designed for the research in molecular sciences under an open-source license.
Data files can be accessed at: Google Drive
Note that the latest version now is v1.1.0.
See details in CHANGELOG.md.
- git clone this repo.
- download and put all
*.pklfiles under./download/raw/directory.
pip install -f requirements.txt
Note that torch_geometric may need to be installed separately follow the instruction on PyG Documentation.
Make sure current work directory is under the repository root
pwd
Then run the data processing script.
sh process_data.sh
Take a cup of coffee☕️. This could take tens of minutes.
Q&A:
- All files are created and tested with
pytho==3.11.9&numpy==2.1.1. - If the repo is cloned into
/abc/QO2Mol, all.pklfiles should be under/abc/QO2Mol/download/raw/. - We offer MD5 hash values in
CHANGELOG.mdfor download check. - If
sh process_data.shcannot locateqo2mol.py, try absolute path with commandpython /abc/QO2Mol/src/dataset/qo2mol/qo2mol.py. - Mac OS may have unexpected issues associated with the file system, Linux is strongly recommended.
Make sure current work directory is under the repository root
pwd
We provide a usage demo that simply requires running the script
sh scripts\dimenetpp\qm_benchmark\tr_baseline_e.sh
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc-sa/4.0/.