ColabFit-Data

The ColabFit Exchange hosts datasets created for the purpose of fitting machine learning interatomic potentials (MLIPs).

Use ColabFit Data-Lake to request inclusion of an existing dataset or submit your own dataset to the ColabFit Exchange.

Overview

To nominate an existing dataset for inclusion in the ColabFit Exchange:

• Open an issue on this repository
• Fill out the information template and submit the issue

To submit your own dataset:

• Open an issue on this repository
• Fill out the information template and submit the issue
• Create a fork of the main branch of this repository
• Upload your dataset files to a single directory on your fork
• Submit a pull request to the main branch of this repository

Once you have created a dataset submission or request, we will correspond with you through the GitHub issue or pull request interface.

Helpful links

ColabFit.org: Explore the ColabFit Exchange

Property definitions from colabfit-tools: Examples, should you need to create your own property definition.

For more information about ColabFit, please contact Prof. Ellad Tadmor and Prof. Stefano Martiniani at support@colabfit.org.