Hugging Face Introduces Community Evals for Transparent Model Benchmarking

Hugging Face has launched Community Evals, a feature that enables benchmark datasets on the Hub to host their own leaderboards and automatically collect evaluation results from model repositories. By Daniel Dominguez