# Open Catalyst 2020 (OC20) IS2RE/IS2RS In-Domain Validation Set, ColabFit Conversion

## Facts

- Provider: Chanussot et al., Meta FAIR and Carnegie Mellon University (Open Catalyst Project); converted by ColabFit Exchange
- License: CC-BY-4.0 (https://raw.githubusercontent.com/facebookresearch/fairchem/main/docs/catalysts/datasets/oc20.md)
- Size: 93 files
- Format: Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/colabfit/OC20_IS2RES_val_id
- Access: public, readable without an account
- DOI: https://doi.org/10.60732/b4005525
- Category: Chemistry & Materials
- Source: https://doi.org/10.1021/acscatal.0c04525
- Page: https://scigantic.com/archive/45ce3003-5f1d-5890-a01c-4589924d3c47
- Keywords: Chemistry & Materials, DFT, Parquet, Hugging Face

## About

ColabFit Exchange conversion of val_id, the in-domain validation split for the Initial Structure to Relaxed Structure (IS2RS) and Initial Structure to Relaxed Energy (IS2RE) tasks of Open Catalyst 2020, published by Chanussot, Das, Goyal, Ulissi, Zitnick and colleagues at Meta FAIR and Carnegie Mellon University (ACS Catalysis 2021). OC20 holds 1,281,040 DFT relaxations (about 265 million single points) of adsorbates (nitrogen, carbon and oxygen chemistries) on inorganic catalyst surfaces. Per the OC20 paper, calculations used VASP with PAW pseudopotentials and the RPBE functional without spin polarization; the per-row method field here reads DFT-rPBE. ColabFit lists 5,024,223 configurations with energy, adsorption energy and atomic forces, plus OC20 mapping metadata such as adsorbate id and symbols, Materials Project bulk id and Miller index. Measured 2026-11-13: ds.parquet plus a co/ folder of 90 parquet shards (17.4 GB); repository storage 45.3 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 validate catalyst machine learning potentials and adsorption energy models (SchNet, DimeNet++, GemNet). Licensed CC BY 4.0; cite Chanussot et al. 2021 and the ColabFit Exchange.

## Load the data

```python
from datasets import load_dataset
ds = load_dataset("colabfit/OC20_IS2RES_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/45ce3003-5f1d-5890-a01c-4589924d3c47
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/45ce3003-5f1d-5890-a01c-4589924d3c47
- Over MCP: `get_schema_card` with id `45ce3003-5f1d-5890-a01c-4589924d3c47` 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)
