metadata

extra_gated_prompt: >-
  You agree to not use the dataset to conduct experiments that cause harm to
  human subjects.
extra_gated_fields:
  Company: text
  Name: text
  Position: text
  Purpose_of_data: text
  Contact_Email: text
  Country: country
  Specific_date: date_picker
  Supervisor_Name (If not applicable, please enter 'None'):
    type: text
  Supervisor_Email (If not applicable, please enter 'None'):
    type: text
  I want to use this dataset for:
    type: select
    options:
      - Research
      - Education
      - label: Other
        value: other
  I agree to use this dataset for non-commercial use ONLY: checkbox
extra_gated_heading: Acknowledge license to accept the repository
extra_gated_description: Our team may take 2–3 days to process your request
extra_gated_button_content: Acknowledge license

Please provide your real name and institutional email for verification.
If you experience any issues, contact Mr. Cao at: 📧 [email protected]
Fake names or unverifiable information will result in your request being denied.

CPPbenchmark

CPPbenchmark is a curated benchmark suite for evaluating machine learning models on crystal property prediction (CPP) tasks. It includes eight tasks—seven regression (e.g., formation energy, band gap, elastic moduli) and one classification (metal/non-metal)—using high-quality datasets derived from the Materials Project. All data is provided in ASE .db format, enabling easy integration with atomistic ML workflows.

📝 Access Request Form

Please fill in the form truthfully. We will review your request within 2–3 business days.

📥 Access & Usage

All required information must be filled out truthfully.
Personal verification typically takes 2 working days.
Once approved, you'll be granted download access to the full database.

📚 Citation & License

Commercial use is strictly prohibited.
All access will be logged.
If you use this dataset in your research, please cite all the following works:

@article{jain2020materials,
  title={The materials project: Accelerating materials design through theory-driven data and tools},
  author={Jain, Anubhav and Montoya, Joseph and Dwaraknath, Shyam and Zimmermann, Nils ER and Dagdelen, John and Horton, Matthew and Huck, Patrick and Winston, Donny and Cholia, Shreyas and Ong, Shyue Ping and others},
  journal={Handbook of Materials Modeling: Methods: Theory and Modeling},
  pages={1751--1784},
  year={2020},
  publisher={Springer}
}

@inproceedings{binsimxrd,
  title={SimXRD-4M: Big Simulated X-ray Diffraction Data and Crystal Symmetry Classification Benchmark},
  author={Bin, CAO and Liu, Yang and Zheng, Zinan and Tan, Ruifeng and Li, Jia and Zhang, Tong-yi},
  booktitle={The Thirteenth International Conference on Learning Representations}
}


@misc{caobin_2025cpp,
    author       = {Bin Cao and Tong-Yi Zhang },
    title        = { CPPbenchmark (Revision 261622f) },
    year         = 2025,
    url          = { https://huggingface.co/datasets/caobin/CPPbenchmark },
    doi          = { 10.57967/hf/5378 },
    publisher    = { Hugging Face }
}

Task Overview

This benchmark includes eight primary tasks for evaluating CPP models in crystal property prediction, divided into regression and classification missions.

Mission Settings

🔢 Regression Tasks

T1: Formation Energy Prediction (float) | eV/atom
T2: Band Gap Prediction (float) | eV
T3: Bulk Modulus Prediction (float) | Gpa
T4: Shear Modulus Prediction (float) | Gpa
T5: Young’s Modulus Prediction (float) | Gpa
T6: Poisson’s Ratio Prediction (float) | NAN
T7: Pugh’s Ratio Prediction (float) | NAN

🧮 Classification Task

T8: Metal/Non-metal Classification (int, binary classification)

Dataset Description

The datasets are hosted on Hugging Face and stored in ASE database (.db) format. Each database includes crystal structures and corresponding property labels.

📁 `fe_bandg/`

Train: MP_100_bgfe_train.db (86,071 crystals)
Val: MP_100_bgfe_val.db (12,295 crystals)
Test: MP_100_bgfe_test.db (24,593 crystals)
Tasks: T1, T2
- Keys: formation_energy, band_gap

📁 `modulus/`

Train: MP_modulus_train.db (6,631 crystals)
Val: MP_modulus_val.db (947 crystals)
Test: MP_modulus_test.db (1,895 crystals)
Tasks: T3–T7
- Keys: bulk_modulus, shear_modulus, youngs_modulus, poissons_ratio, pughs_modulus_ratio

📁 `metal_nometal/`

Train: MP_100_metal_train.db (86,071 crystals)
Val: MP_100_metal_val.db (12,295 crystals)
Test: MP_100_metal_test.db (24,593 crystals)
Task: T8 (label: 0 or 1)
- Keys: metal

Data Format

All data is stored in ASE database format (.db). Each entry contains both the crystal structure and the associated target properties.

🧪 Example: Reading Data from an ASE Database

You can read data using the ase.db module as follows:

from ase.db import connect

# Load the database
db = connect('MP_100_bgfe_train.db')

# Iterate through entries
for row in db.select():
    atoms = row.toatoms()  # ASE Atoms object
    formation_energy = row.formation_energy
    band_gap = row.band_gap

    print(f'Formula: {atoms.get_chemical_formula()}')
    print(f'Formation Energy: {formation_energy:.3f} eV')
    print(f'Band Gap: {band_gap:.3f} eV')
    break  # remove this if you want to iterate over all entries