# GuacaMol Benchmark Molecule Sets (ChEMBL 24)

## Facts

- Provider: Brown, Fiscato, Segler and Vaucher (BenevolentAI); uploaded by katielink
- License: CC-BY-SA-3.0 (https://ftp.ebi.ac.uk/pub/databases/chembl/ChEMBLdb/latest/LICENSE)
- Size: 5 files
- Stored at: Hugging Face, https://huggingface.co/datasets/katielink/GuacaMol
- Access: public, readable without an account
- DOI: https://doi.org/10.6084/m9.figshare.7322252.v2
- Category: Drug Discovery
- Source: https://doi.org/10.1021/acs.jcim.8b00839
- Page: https://scigantic.com/archive/0b4d1b73-42d2-586e-97c7-97529a481bf0
- Keywords: Drug Discovery, benchmark, Hugging Face

## About

Copy of the standard GuacaMol training, validation and test SMILES sets from Benchmarking Models for De Novo Molecular Design (Brown, Fiscato, Segler and Vaucher, Journal of Chemical Information and Modeling 2019). GuacaMol is a widely used benchmark suite for generative chemistry models, with distribution-learning metrics (validity, uniqueness, novelty, Frechet ChemNet Distance, KL divergence) and goal-directed tasks such as rediscovery, similarity, isomer and multi-property optimisation. The authors published the pre-built molecule sets on figshare as a subset of molecules extracted from ChEMBL 24, and the GuacaMol code expects models to be trained on them. Measured 2026-11-14: three plain-text files with one SMILES string per line, guacamol_v1_train.smiles (62 MB), guacamol_v1_valid.smiles (4 MB) and guacamol_v1_test.smiles (12 MB), about 77 MB in total. Read them with pandas.read_csv(header=None), plain Python, or RDKit; the guacamol Python package scores models against them. Use for training SMILES or graph generative models, VAEs, and reinforcement learning molecule generators, and for comparing against the GuacaMol leaderboard. The figshare release is CC BY 4.0 and the source ChEMBL data are CC BY-SA 3.0, so share-alike attribution applies; cite Brown et al. 2019 and ChEMBL.

## Load the data

```python
from huggingface_hub import hf_hub_download
path = hf_hub_download(repo_id="katielink/GuacaMol", filename="guacamol_v1_test.smiles", repo_type="dataset")
print(path)
```

## Data structure

Example file: `guacamol_v1_test.smiles`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/0b4d1b73-42d2-586e-97c7-97529a481bf0
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/0b4d1b73-42d2-586e-97c7-97529a481bf0
- Over MCP: `get_schema_card` with id `0b4d1b73-42d2-586e-97c7-97529a481bf0` at https://api.scigantic.com/mcp

## Related topics

- [Benchmark datasets](https://scigantic.com/topics/benchmark)
- [Drug Discovery datasets](https://scigantic.com/topics/drug-discovery)
