PubChemQCR: Relaxation Trajectories of 3.5 Million Molecules for MLIP Benchmarking
- Provider
- TAMU DIVE Lab (Cong Fu, Shuiwang Ji et al.) with Maho Nakata (RIKEN)
- License
- CC-BY-4.0 (source)
- Size
- 25 files
- Stored at
- Hugging Face (https://huggingface.co/datasets/divelab/PubChemQCR)
- Access
- Free and public. No account with the data provider is needed.
- Category
- Chemistry & Materials
- Source
- https://arxiv.org/abs/2506.23008
About this dataset
PubChemQCR, from the TAMU DIVE Lab (Shuiwang Ji and colleagues, with Maho Nakata of RIKEN), contains the geometry relaxation trajectories of about 3.5 million small molecules from the PubChemQC project, optimised sequentially with PM3, Hartree-Fock and DFT, for a total that the authors put at about 300 million molecular conformations with energies and forces. It is released with a benchmark for machine learning interatomic potentials on quantum chemistry relaxations (arXiv:2506.23008). Measured 2026-11-14, the repository root holds LMDB shards Data01.lmdb to Data12.lmdb plus test.lmdb (most around 50 GB each), a Code/ folder with a loader and LMDB structure notes, and uses about 430 GB in total. The loader selects an optimisation stage (pm3, hf, first and second DFT substages, or mixing) and whether to return whole trajectories; read the LMDB files with the python lmdb package and the provided code. The card states that, authorised by Maho Nakata, the processed data are released by TAMU DIVE Lab under CC-BY-4.0.
How to load this dataset in Python
from huggingface_hub import hf_hub_download
path = hf_hub_download(repo_id="divelab/PubChemQCR", filename="Code/data.py", repo_type="dataset")
print(path)Data structure
Files
Example file: Code/data.py
Analyze this dataset without downloading it
On Scigantic, PubChemQCR: Relaxation Trajectories of 3.5 Million Molecules for MLIP Benchmarking opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.
This page is also available as Markdown and as JSON.
Related Chemistry & Materials datasets
- Nemotron Science V 1
- NequIP Benchmark Trajectories: LiPS, Li4P2O7 Melt-Quench and Formate on Cu, ColabFit Conversion
- AIMD-Chig Chignolin Ab Initio MD, Random Split Training Set, ColabFit Conversion
- Amorphous and Liquid Carbon Melt-Quench AIMD Dataset (Minamitani et al. 2023), ColabFit Conversion
- GAP-20 Carbon Potential Training Data: Crystalline, Amorphous, Surfaces and Defects at optB88-vdW DFT (Rowe et al. 2020, Cambridge Apollo), ColabFit Conversion
- Prism Edge Dislocations in Mg and Mg-Zn Alloys: Solute Strengthening DFT (Niazi and Curtin 2024, Materials Cloud), ColabFit Conversion
- JARVIS ALIGNN-FF Training Set: OptB88vdW Energies, Forces and Stresses for 89 Elements, ColabFit Conversion
- Flat Bands in Graphene Superlattices with Transition-Metal Adatoms: DFT Structures (Skurativska et al. 2021, Materials Cloud), ColabFit Conversion