# MAD-1.5 Massive Atomic Diversity PBE Companion Set, ColabFit Conversion

## Facts

- Provider: Cesare Malosso, Michele Ceriotti, Arslan Mazitov and coauthors (EPFL COSMO); ColabFit conversion
- License: CC-BY-4.0 (https://archive.materialscloud.org/api/records/18tke-tt476)
- Size: 6 files
- Format: Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/colabfit/Massive_Atomic_Diversity_MAD-1.5_PBE
- Access: public, readable without an account
- DOI: https://doi.org/10.24435/materialscloud:jc-9f
- Category: Chemistry & Materials
- Source: https://doi.org/10.48550/arXiv.2603.02089
- Page: https://scigantic.com/archive/16c029e9-fc60-5cb0-aaaa-db1a7f944b2a
- Keywords: Chemistry & Materials, Parquet, Hugging Face

## About

PBE recomputation of the MAD-1 subsets of the Massive Atomic Diversity 1.5 dataset (MC3D bulk crystals and their rattled, random, surface and cluster variants, MC2D two-dimensional materials, SHIFTML molecular crystals and fragments) plus monomers and extended random structures, with FHI-aims all-electron settings matched to the r2SCAN set. It was used as a second prediction head when training PET-MAD-1.5 and improved force accuracy by about 25 percent over r2SCAN-only training; the file combines the train, validation and test splits. ColabFit counts 101,493 configurations (about 2.6 million atoms over 102 elements) with energy, atomization energy, forces and stress, for multi-fidelity PBE versus r2SCAN studies. This is the ColabFit Exchange conversion: configurations sit in Parquet files under co/ (positions, cell, species, 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 schema at materials.colabfit.org/docs. Published on Materials Cloud under CC-BY-4.0; cite Malosso et al., arXiv:2603.02089 (2026).

## Load the data

```python
from datasets import load_dataset
ds = load_dataset("colabfit/Massive_Atomic_Diversity_MAD-1.5_PBE")
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/16c029e9-fc60-5cb0-aaaa-db1a7f944b2a
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/16c029e9-fc60-5cb0-aaaa-db1a7f944b2a
- Over MCP: `get_schema_card` with id `16c029e9-fc60-5cb0-aaaa-db1a7f944b2a` 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)
