Hugging Face has announced a major update to its Hub platform: the introduction of pull requests and discussions. These new features, available under the Community tab for all repository types including models, datasets, and Spaces, enable open collaboration among community members. Users can now propose changes, ask questions, and provide feedback directly on repositories, streamlining the process of improving machine learning projects.
Discussions allow users to start conversations around repositories, while pull requests enable direct contributions without the need for forking. Instead, contributors push changes to special ref branches on the original repository, simplifying the workflow. The goal is to create a simpler, ML-focused alternative to traditional git hosts like GitHub, with no hard distinction between issues and PRs.
The update aligns with ethical ML practices by centralizing feedback and iteration. Example use cases include suggesting improvements to model cards to address biases, flagging concerning outputs in demos, or adding TensorFlow weights to existing models.
Hugging Face plans to continue refining these features based on community input. Users are encouraged to join discussions and open pull requests to start collaborating.