HF-NLP10K: A Dataset and Metadata Analysis of 10,000+ Hugging Face NLP Models

Hugging Face has become a central platform for sharing NLP models. However, inconsistent and incomplete metadata makes it difficult to find suitable models. In this project, we introduce HF-NLP10K, a structured dataset of over 10,000 NLP models, including many LLMs, enriched with 33 metadata fields, combining Hugging Face API data with LLM-based parsing of free-text model cards. Here we provide: 📦 The HF-NLP10K dataset (CMA-PROJ_Dataset.csv.gz) 📜 Scripts for dataset creation Primary contact: Michael Färber (TU Dresden and ScaDS.AI, Germany) – [e-mail]