Hugging Face Model Card Guidelines

This document is the Hugging Face Hub documentation page for Model Cards, a standardized metadata and documentation framework that the platform requires or recommends for models hosted on the Hub. The documentation states that model cards are files that accompany models and provide information about training data, evaluation results, intended uses, limitations, and bias disclosures, and the Hub uses structured YAML metadata in model card headers to enable filtering, discovery, and dataset/evaluation linkage. The document establishes that model card metadata fields including license, language, tags, datasets, and metrics are parsed by the Hub to power search and filtering infrastructure, making metadata completeness operationally significant for model discoverability and downstream use attribution. The documentation engages with responsible AI disclosure norms referenced in frameworks such as the EU AI Act, which includes transparency and documentation obligations for high-risk AI systems, and model cards as described here provide a disclosure mechanism that may interact with those requirements depending on the model type and deployment context. Compliance teams reviewing AI system documentation obligations under the EU AI Act or similar frameworks should evaluate whether model card disclosures as structured here satisfy applicable transparency, intended-use documentation, and limitation-disclosure requirements for their specific model category and jurisdiction.