Huatuo-26M is a massive, 26 million-pair Chinese medical question-and-answer dataset designed for AI research in healthcare. It enables the development of advanced NLP applications, machine learning models for medical tasks, and intelligent medical systems, catering to researchers and developers in the medical AI domain.

How It Works

The dataset aggregates Q&A pairs from diverse sources, including online medical encyclopedias, knowledge bases, and consultation records. A refined version, Huatuo-Lite, offers enhanced data quality and additional fields like hospital departments and related diseases. This multi-source approach ensures broad coverage of medical topics, from diseases and symptoms to treatments and drug information.

Quick Start & Requirements

• Install via Hugging Face datasets: datasets.load_dataset('FreedomIntelligence/huatuo_knowledge_graph_qa') (and similar for other sub-datasets).
• Requires Python and the datasets library.
• Links: Huatuo-Lite, huatuo_encyclopedia_qa, huatuo_knowledge_graph_qa, huatuo_consultation_qa, huatuo26M-testdatasets.

Highlighted Details

• Over 26 million high-quality Chinese medical Q&A pairs.
• Covers diseases, symptoms, treatments, and drug information.
• Huatuo-Lite offers refined data with additional fields (department, related diseases).
• Usable for Q&A systems, text classification, sentiment analysis, disease prediction, and LLM fine-tuning.

Maintenance & Community

• Contact: xidongw@163.com or via GitHub Issues.
• Citation available for academic use.

Licensing & Compatibility

• Licensed under Apache 2.0.
• Permissive for commercial use and integration with closed-source projects.

Limitations & Caveats

The huatuo_consultation_qa dataset primarily contains URLs as answers, requiring further processing to extract actionable information. The dataset is primarily focused on Chinese medical information.