Published May 5, 2025 | Version nc_1000_v1.1
Dataset Open

modelforge curated dataset: ANI-2x

Description

Modelforge Curated ANI-2x Dataset:
- 1000 configuration test set
Version: nc_1000_v1.1

This provides a curated hdf5 file for a subset of the ANI-2x dataset designed to be compatible with modelforge, an infrastructure to implement and train NNPs.  This dataset contains 101 unique records for 1000  total configurations, with a maximum of 10 per system.   Note, configurations are parititioned into records based on the array of atomic species appearing in sequence in the source data file.

When applicable, the units of properties are provided in the datafile,  encoded as strings compatible with the openff-units package.  For more information about the structure of the data file, please see the following:

This dataset is compatible with modelforge hdf5 schema 2.

Properties Included:     

  • atomic_numbers 
  • positions       
    • "per_atom"
    •  "nanometer"
  • forces  
    • "per_atom"
    •  "kilojoule_per_mole / nanometer"
  • energies        
    • "per_system"
    • "kilojoule_per_mole"

Source Dataset:

The ANI-2x data set includes properties for small organic molecules that contain H, C, N, O, S, F, and Cl. This dataset contains 9651712 conformers. This data was generated with the wB97X/631Gd level of theory used in the original ANI-2x paper, calculated using Gaussian 09.

Citations:

ANI-2x publication:

  • Devereux, C, Zubatyuk, R., Smith, J. et al. "Extending the applicability of the ANI deep learning molecular potential to sulfur and halogens." Journal of Chemical Theory and Computation 16.7 (2020): 4192-4202. https://doi.org/10.1021/acs.jctc.0c00121

Source dataset, released with CC Attribution 4.0 International license:

Files

Files (179.3 kB)

Name Size Download all
md5:574cb77f5089019965607f7c7110d4c2
179.3 kB Download

Additional details

Related works

Is derived from
Dataset: 10.5281/zenodo.10108942 (DOI)
Is described by
Journal article: 10.1021/acs.jctc.0c00121 (DOI)

Software

Repository URL
https://github.com/choderalab/modelforge
Programming language
Python
Development Status
Active