Skip to main content

Hugging Face Repositories

Note

Support for Hugging Face Xet is coming soon to Sonatype Nexus Repository.

Hugging Face is a machine learning platform and community best known for its extensive library of pre-trained transformer models for various tasks, such as natural language processing, computer vision, and audio analysis. The platform provides tools for building, training, and deploying machine learning models and datasets, and spaces to share and collaborate with others.

Hugging Face platform consists of the following components:

  • Models: Core components of Hugging Face, pre-trained brains to perform tasks. They are the recipes that have already been perfected.

  • Datasets: Collections of data used to train and evaluate the models. They are the ingredients for the recipes.

  • Spaces: They are like interactive demos or apps built with Hugging Face models. You can use them to test out models to see how they work.

Nexus Repository 3 supports proxying Hugging Face models and datasets in both cloud and self-hosted (Pro and CE) deployments.

Getting Started with Hugging Face in Nexus

Explore the following topics to learn how to configure and use Hugging Face with Nexus Repository.

  1. Nexus for Hugging Face

  2. Create a Hugging Face Repository

  3. Configure Hugging Face with Nexus

  4. Hugging Face CLI Usage

Nexus for Hugging Face

  • Proxy Repository: Centralize access to Hugging Face models through a Nexus proxy repository.

  • Manage access through Nexus authentication and authorization controls.

  • Supports Bearer token authentication starting with Nexus Repository 3.95.

  • Retain Hugging Face Git commit hashes when caching different repository revisions.