# ANI-1x Active Learning Molecular Conformers (wB97x/6-31G*), ColabFit Conversion

## Facts

- Provider: Smith, Zubatyuk, Nebgen, Lubbers, Barros, Roitberg, Isayev and Tretiak (University of Florida, Los Alamos National Laboratory, Carnegie Mellon University); converted by ColabFit Exchange
- License: CC-BY-4.0 (https://api.figshare.com/v2/articles/10047041)
- Size: 4 files
- Format: Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/colabfit/ANI-1x
- Access: public, readable without an account
- DOI: https://doi.org/10.60732/dd0270c8
- Category: Chemistry & Materials
- Source: https://doi.org/10.1038/s41597-020-0473-z
- Page: https://scigantic.com/archive/71ad430f-5a5c-53d5-aee7-feab4fdc166f
- Keywords: Chemistry & Materials, molecular dynamics, quantum chemistry, interatomic potentials, Parquet, Hugging Face

## About

ColabFit Exchange conversion of ANI-1x, the active learning dataset published by Justin S. Smith, Olexandr Isayev, Sergei Tretiak, Adrian E. Roitberg and colleagues (Scientific Data 2020). Starting from an initial training set, conformations were added iteratively wherever an ensemble of ANI models disagreed; molecules were drawn from GDB-11, ChEMBL and other sources and sampled with molecular dynamics, normal mode, dimer and torsion sampling. The upstream HDF5 release (about 5 million conformations) provides wB97x/6-31G*, wB97x/def2-TZVPP and CCSD(T)*/CBS data; the rows in this conversion carry software Gaussian 09, method wB97x and basis set 6-31G*. ColabFit lists 308,645 configurations of C, H, N and O molecules with energy and atomic forces, so treat it as a subset of the full release. Measured 2026-11-13: ds.parquet plus a co/ folder holding a single 319 MB parquet file; repository storage 1.4 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. Use it to train and test ANI-style and equivariant machine learning interatomic potentials for organic molecules. The upstream figshare release is CC0 and ColabFit distributes this copy under CC BY 4.0; cite Smith et al. 2020 and the ColabFit Exchange.

## Load the data

```python
from datasets import load_dataset
ds = load_dataset("colabfit/ANI-1x")
print(ds)
```

## Data structure

File types in sampled files: Parquet

Example file: `co/co_0.parquet`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/71ad430f-5a5c-53d5-aee7-feab4fdc166f
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/71ad430f-5a5c-53d5-aee7-feab4fdc166f
- Over MCP: `get_schema_card` with id `71ad430f-5a5c-53d5-aee7-feab4fdc166f` at https://api.scigantic.com/mcp

## Related topics

- [Parquet datasets](https://scigantic.com/topics/parquet)
- [Chemistry & Materials datasets](https://scigantic.com/topics/chemistry-and-materials)
- [Interatomic potentials datasets](https://scigantic.com/topics/interatomic-potentials)
- [Molecular dynamics datasets](https://scigantic.com/topics/molecular-dynamics)
- [Quantum chemistry datasets](https://scigantic.com/topics/quantum-chemistry)
