Hugging Face has introduced a new feature allowing users to generate Digital Object Identifiers (DOIs) for their models and datasets directly from the Hub. This initiative aims to enhance reproducibility, documentation, and ease of sharing within the machine learning community.
DOIs are unique strings that identify digital objects, such as datasets and models, and are tied to metadata like URL, version, and creation date. They serve as permanent references, similar to an ISBN for books, ensuring resources remain findable and citable over time.
Users can request a DOI through a partnership with DataCite by filling out required metadata in their repository settings. Once generated, others can cite the work by clicking "Cite this model/dataset" on the model or dataset page. DOIs can be updated for new versions, with previous versions marked as outdated.
This feature was developed in response to community feedback. Hugging Face encourages further suggestions via Twitter or GitHub issues.