# OrbNet Denali Training Set (wB97X-D3/def2-TZVP), ColabFit Conversion

## Facts

- Provider: Christensen, Sirumalla, Qiao, O'Connor, Smith, Ding, Bygrave, Anandkumar, Welborn, Manby and Miller (Entos Inc. and Caltech); converted by ColabFit Exchange
- License: CC-BY-4.0 (https://api.figshare.com/v2/articles/14883867)
- Size: 6 files
- Format: Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/colabfit/OrbNet_Denali
- Access: public, readable without an account
- DOI: https://doi.org/10.60732/b6043e2b
- Category: Chemistry & Materials
- Source: https://doi.org/10.1063/5.0061990
- Page: https://scigantic.com/archive/d83de299-e581-54aa-8d60-8f8fb38a591a
- Keywords: Chemistry & Materials, DFT, quantum chemistry, benchmark, Parquet, Hugging Face

## About

ColabFit Exchange conversion of the OrbNet Denali training set, published by Anders S. Christensen, Thomas F. Miller III, Frederick R. Manby and colleagues at Entos and Caltech (Journal of Chemical Physics 2021) to train OrbNet Denali, a machine learning model built on GFN-xTB semiempirical quantum chemistry features. Per the upstream description, all DFT single-point calculations were run in Entos Qcore 0.8.17 at the wB97X-D3/def2-TZVP level of theory with in-core density fitting. Elements are B, Br, C, Ca, Cl, F, H, I, K, Li, Mg, N, Na, O, P, S and Si. ColabFit lists 2,337,230 configurations (104.9 million atoms) with energy only. Measured 2026-11-13: ds.parquet plus a co/ folder of 3 parquet files (3.3 GB); repository storage 6.9 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. Use it to train or benchmark machine learning potentials and delta-learning models for conformer energies, tautomers, protonation states and noncovalent complexes. Licensed CC BY 4.0; cite Christensen et al. 2021 and the ColabFit Exchange.

## Load the data

```python
from datasets import load_dataset
ds = load_dataset("colabfit/OrbNet_Denali")
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/d83de299-e581-54aa-8d60-8f8fb38a591a
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/d83de299-e581-54aa-8d60-8f8fb38a591a
- Over MCP: `get_schema_card` with id `d83de299-e581-54aa-8d60-8f8fb38a591a` at https://api.scigantic.com/mcp

## Related topics

- [Parquet datasets](https://scigantic.com/topics/parquet)
- [Benchmark datasets](https://scigantic.com/topics/benchmark)
- [Chemistry & Materials datasets](https://scigantic.com/topics/chemistry-and-materials)
- [DFT datasets](https://scigantic.com/topics/dft)
- [Quantum chemistry datasets](https://scigantic.com/topics/quantum-chemistry)
