Name: Massive Atomic Diversity MAD bench mptrj
License: CC-BY-4.0

Dataset

Massive_Atomic_Diversity_MAD_bench_mptrj

Download Original Data Files 1.9 MB

Download Dataset Parquet Files 3.5 MB

Download Dataset XYZ Files 2.1 MB

Find on Hugging Face 🤗

View the MAD family of datasets

Species content of dataset

Name :

Massive_Atomic_Diversity_MAD_bench_mptrj

Authors :

Arslan Mazitov, Sofiia Chorna, Guillaume Fraux, Marnik Bercx, Giovanni Pizzi, Sandip De, Michele Ceriotti

Description :

The MAD benchmark dataset, containing a selection of MAD test, MPtrj, Alexandria, SPICE, MD22 and OC2020 datasets, computed with MPtrj DFT settings. Part of the MAD (Massive Atomic Diversity) dataset family. From the creators: Starting from relatively small sets of stable structures, the dataset is built to contain “massive atomic diversity” (MAD) by aggressively distorting these configurations, with near-complete disregard for the stability of the resulting configurations. The electronic structure details, on the other hand, are chosen to maximize consistency rather than to obtain the most accurate prediction fora given structure, or to minimize computational effort. The MAD dataset we present here, despite containing fewer than 100k structures, has already been shown to enable training universal interatomic potentials that are competitive with models trained on traditional datasets with two to three orders of magnitude more structures.

DOI :

https://doi.org/10.60732/30653c33 https://commons.datacite.org/doi.org/10.60732/30653c33

Cite As :

Mazitov, A., Chorna, S., Fraux, G., Bercx, M., Pizzi, G., De, S., and Ceriotti, M. "Massive Atomic Diversity MAD bench mptrj." ColabFit, 2025. https://doi.org/10.60732/30653c33.

ColabFit ID :

DS_x44bl0q0dyxr_0

Extended ID :

Massive_Atomic_Diversity_MAD_bench_mptrj__Mazitov-Chorna-Fraux-Bercx-Pizzi-De-Ceriotti__DS_x44bl0q0dyxr_0

Date Added :

2025-08-05

License :

CC-BY-4.0

Downloads :

Num. Configurations :

2,114

Num. Atoms :

58,755

Calculated Property Types :

atomic_forces cauchy_stress energy

Elements :

Ag (0.53%) Al (1.26%) As (0.45%) Au (0.46%) B (0.6%) Ba (0.36%) Be (0.11%) Bi (0.22%) Br (0.56%) C (30.49%) Ca (0.82%) Cd (0.5%) Ce (0.16%) Cl (0.89%) Co (0.54%) Cr (0.21%) Cs (0.25%) Cu (0.7%) Dy (0.29%) Er (0.35%) Eu (0.05%) F (1.21%) Fe (0.55%) Ga (0.78%) Gd (0.03%) Ge (0.51%) H (18.65%) He (0.0%) Hf (0.69%) Hg (0.31%) Ho (0.27%) I (0.48%) In (1.08%) Ir (0.51%) K (0.55%) Kr (0.0%) La (0.23%) Li (0.54%) Lu (0.11%) Mg (0.77%) Mn (0.4%) Mo (0.21%) N (4.86%) Na (0.64%) Nb (0.28%) Nd (0.2%) Ni (0.67%) Np (0.02%) O (8.83%) Os (0.25%) P (0.93%) Pb (0.31%) Pd (0.86%) Pm (0.04%) Pr (0.23%) Pt (0.68%) Pu (0.0%) Rb (0.2%) Re (0.35%) Rh (0.77%) Ru (0.58%) S (1.67%) Sb (0.77%) Sc (0.37%) Se (1.05%) Si (0.89%) Sm (0.2%) Sn (0.78%) Sr (0.72%) Ta (0.34%) Tb (0.1%) Tc (0.16%) Te (0.59%) Th (0.01%) Ti (0.81%) Tl (0.49%) Tm (0.22%) U (0.02%) V (0.39%) W (0.09%) Xe (0.01%) Y (0.62%) Yb (0.14%) Zn (0.7%) Zr (0.47%)

Methods :

DFT-PBEsol

Software :

VASP

Publication Link :

https://doi.org/10.48550/arXiv.2506.19674

Data Source Link :

https://doi.org/10.24435/materialscloud:vd-e8

Spec File :

colabfitspec.json

Configuration Sets by Name :

Configuration Sets by ID :

Name: Massive_Atomic_Diversity_MAD_bench_mptrj Extended ID: Massive_Atomic_Diversity_MAD_bench_mptrj__Mazitov-Chorna-Fraux-Bercx-Pizzi-De-Ceriotti__DS_x44bl0q0dyxr_0 Description: The MAD benchmark dataset, containing a selection of MAD test, MPtrj, Alexandria, SPICE, MD22 and OC2020 datasets, computed with MPtrj DFT settings. Part of the MAD (Massive Atomic Diversity) dataset family. From the creators: Starting from relatively small sets of stable structures, the dataset is built to contain “massive atomic diversity” (MAD) by aggressively distorting these configurations, with near-complete disregard for the stability of the resulting configurations. The electronic structure details, on the other hand, are chosen to maximize consistency rather than to obtain the most accurate prediction fora given structure, or to minimize computational effort. The MAD dataset we present here, despite containing fewer than 100k structures, has already been shown to enable training universal interatomic potentials that are competitive with models trained on traditional datasets with two to three orders of magnitude more structures. Authors: Arslan Mazitov Sofiia Chorna Guillaume Fraux Marnik Bercx Giovanni Pizzi Sandip De Michele Ceriotti DOI: 10.60732/30653c33 Calculated Property Types: atomic_forces cauchy_stress energy Elements: Ag (0.53%) Al (1.26%) As (0.45%) Au (0.46%) B (0.6%) Ba (0.36%) Be (0.11%) Bi (0.22%) Br (0.56%) C (30.49%) Ca (0.82%) Cd (0.5%) Ce (0.16%) Cl (0.89%) Co (0.54%) Cr (0.21%) Cs (0.25%) Cu (0.7%) Dy (0.29%) Er (0.35%) Eu (0.05%) F (1.21%) Fe (0.55%) Ga (0.78%) Gd (0.03%) Ge (0.51%) H (18.65%) He (0.0%) Hf (0.69%) Hg (0.31%) Ho (0.27%) I (0.48%) In (1.08%) Ir (0.51%) K (0.55%) Kr (0.0%) La (0.23%) Li (0.54%) Lu (0.11%) Mg (0.77%) Mn (0.4%) Mo (0.21%) N (4.86%) Na (0.64%) Nb (0.28%) Nd (0.2%) Ni (0.67%) Np (0.02%) O (8.83%) Os (0.25%) P (0.93%) Pb (0.31%) Pd (0.86%) Pm (0.04%) Pr (0.23%) Pt (0.68%) Pu (0.0%) Rb (0.2%) Re (0.35%) Rh (0.77%) Ru (0.58%) S (1.67%) Sb (0.77%) Sc (0.37%) Se (1.05%) Si (0.89%) Sm (0.2%) Sn (0.78%) Sr (0.72%) Ta (0.34%) Tb (0.1%) Tc (0.16%) Te (0.59%) Th (0.01%) Ti (0.81%) Tl (0.49%) Tm (0.22%) U (0.02%) V (0.39%) W (0.09%) Xe (0.01%) Y (0.62%) Yb (0.14%) Zn (0.7%) Zr (0.47%) Methods: DFT-PBEsol Software: VASP Number of Configurations: 2,114 Number of Atoms: 58,755 Publication Link: https://doi.org/10.48550/arXiv.2506.19674 Data Source Link: https://doi.org/10.24435/materialscloud:vd-e8

Dataset viewer powered by Hugging Face

Report an issue with this dataset

Related Datasets

No uploaded content is transferred in ownership from the original creators to ColabFit. All content is distributed under the license specified by its contributor who has stated that he or she has the authority to share it under the specified license.