ANI-1: 20 Million DFT Energies for Non-Equilibrium Small Molecules, ColabFit Conversion
- Provider
- Smith, Isayev and Roitberg (University of Florida, University of North Carolina at Chapel Hill); converted by ColabFit Exchange
- License
- CC-BY-4.0 (source)
- Size
- 108 files
- Format
- Parquet (in sampled files)
- Stored at
- Hugging Face (https://huggingface.co/datasets/colabfit/ANI-1)
- Access
- Free and public. No account with the data provider is needed.
- DOI
- 10.60732/a57b3cb3
- Category
- Chemistry & Materials
- Source
- https://doi.org/10.1038/sdata.2017.193
About this dataset
ColabFit Exchange conversion of ANI-1, the dataset of more than 20 million non-equilibrium conformations published by Justin S. Smith, Olexandr Isayev and Adrian E. Roitberg (Scientific Data 2017) that trained the original ANI-1 neural network potential. Conformations were generated by normal mode sampling of a subset of GDB-11 molecules with 1 to 8 heavy atoms limited to C, N and O (plus hydrogen); the upstream release states that DFT energies were computed with the wB97x functional and the 6-31G(d) basis set. ColabFit lists 24,389,594 configurations with energy only (no forces), and defines configuration sets separating standard and high energy conformers (more than 275 kcal/mol above the lowest energy conformer) and, within those, the number of heavy atoms. Measured 2026-11-13: ds.parquet, a co/ folder of 103 parquet shards (13.3 GB), and cs/ and cs_co_map/ folders for the configuration sets; repository storage 15.9 GB. Rows follow the ColabFit configuration schema (positions, atomic_numbers, cell, pbc, energy, atomic_forces, cauchy_stress, software, method and per-row metadata) and open with pandas.read_parquet, pyarrow or datasets.load_dataset; ds.parquet holds the dataset summary. It is a standard benchmark for energy-only machine learning potentials and conformer energy models in computational chemistry. The upstream figshare release is CC0 and ColabFit distributes this copy under CC BY 4.0; cite Smith et al. 2017 and the ColabFit Exchange.
How to load this dataset in Python
from datasets import load_dataset
ds = load_dataset("colabfit/ANI-1")
print(ds)Data structure
File types in sampled files: Parquet
Files
Example file: co/co_0.parquet
Analyze this dataset without downloading it
On Scigantic, ANI-1: 20 Million DFT Energies for Non-Equilibrium Small Molecules, ColabFit Conversion 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
- Solvated Protein Fragments (PhysNet Benchmark, revPBE-D3), ColabFit Conversion
- Acetylacetone (acac) Proton-Transfer Benchmark: 300 K MD Training Set with PBE-D3 Forces (BOTNet, Batatia et al.), ColabFit Conversion
- Plate reader instrument output with ASM: 140 exports from Agilent Gen5, Agilent Gen5 Image, BMG Labtech SMART Control, BMG Labtech MARS and 10 more
- QM-22 Malonaldehyde (C3H4O2): Ab Initio Potential Energy Surface Data (Bowman group), ColabFit Conversion
- Nemotron RL Knowledge Web Search Mcqa
- Linear Magnetoelectric Coefficient of Chromia (Cr2O3): DFT under Electric Field in Elk, VASP and Quantum ESPRESSO (2024, Materials Cloud), ColabFit Conversion
- OrbNet Denali Training Set (wB97X-D3/def2-TZVP), ColabFit Conversion
- JARVIS-ML CFID Descriptors for 460K OQMD Crystals: DFT Formation Energies (NIST JARVIS, figshare), ColabFit Conversion