Hugging Face Classrooms

Overview

Hugging Face Classrooms provides teachers and students with a free collaborative workspace on the Hugging Face Hub. An instructor creates a classroom, which is backed by a Hub organization, then invites students to build and share machine learning projects in one place.

What Is Free

• A classroom organization on the Hub, created at no cost from the Classrooms page.
• Free hosting for the classroom's datasets and models, and for Static Spaces. The Classrooms page still says demos are "hosted for free", but the Spaces docs now say Gradio and Docker Spaces "require a paid plan to create" (Team or Enterprise for organizations) and list no exception for classrooms.
• Git-based version control for every model, dataset and demo in the classroom.
• Student access controls with admin, read and write permissions.
• Private or public visibility for classroom content, chosen by the instructor.
• Hugging Face's free courses (NLP, Deep RL, ML demos) and the educator toolkit are available alongside the classroom to support teaching.

How to Get Started

1. Create a free account at huggingface.co.
2. Open huggingface.co/classrooms and create a classroom.
3. Set the classroom name, visibility and member permissions.
4. Invite students by Hugging Face username or email address.

Students can then upload datasets and models, train, fine-tune, experiment and deploy directly from the shared workspace.