# Open Catalyst 2020 (OC20) S2EF In-Domain Validation Set, ColabFit Conversion

## Facts

- Provider: Meta FAIR and Carnegie Mellon University (Chanussot et al.); ColabFit conversion
- License: CC-BY-4.0 (https://raw.githubusercontent.com/facebookresearch/fairchem/main/docs/catalysts/datasets/oc20.md)
- Size: 5 files
- Format: Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/colabfit/OC20_S2EF_val_id
- Access: public, readable without an account
- DOI: https://doi.org/10.60732/eaea5062
- Category: Chemistry & Materials
- Source: https://doi.org/10.1021/acscatal.0c04525
- Page: https://scigantic.com/archive/1a702f9a-2b6f-54f6-a1c5-897fc443ecb4
- Keywords: Chemistry & Materials, DFT, Parquet, Hugging Face

## About

The in-domain validation split of the Open Catalyst 2020 (OC20) Structure to Energy and Forces task, used to report energy and force mean absolute error for machine learning potentials on adsorbate and catalyst surface systems drawn from the same distribution as training. Labels are DFT (RPBE) energies, adsorption energies and per-atom forces, with OC20 mapping metadata (adsorbate id, bulk id, Miller index). ColabFit's card counts 999,866 configurations and about 73 million atoms; the repository uses about 12.2 GB (measured 2026-11-14). The out-of-domain splits (unseen adsorbates, catalysts, or both) are separate ColabFit repositories. This is the ColabFit Exchange conversion: configurations are Parquet files under co/ (structure, positions, cell, computed properties and metadata per row) with dataset-level information in ds.parquet; read them with pyarrow, pandas, polars or Hugging Face datasets, following the ColabFit configuration schema at materials.colabfit.org/docs. Upstream OC20 is CC-BY-4.0 from Meta FAIR; cite Chanussot et al., ACS Catalysis 2021.

## Load the data

```python
from datasets import load_dataset
ds = load_dataset("colabfit/OC20_S2EF_val_id")
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/1a702f9a-2b6f-54f6-a1c5-897fc443ecb4
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/1a702f9a-2b6f-54f6-a1c5-897fc443ecb4
- Over MCP: `get_schema_card` with id `1a702f9a-2b6f-54f6-a1c5-897fc443ecb4` 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)
- [DFT datasets](https://scigantic.com/topics/dft)
