Weikaijie/HealthPulse-Qwen2.5-7B
Weikaijie/HealthPulse-Qwen2.5-7B is a 7.6 billion parameter medical large language model, fine-tuned from Qwen2.5-7B-Instruct using SFT + LoRA on the Huatuo Chinese doctor-patient dataset. It focuses on medical accuracy and safety, excelling in symptom consultation, health education, guided diagnosis, and medication common sense Q&A. The model is optimized for Chinese medical applications with a context length of 32768 tokens.
Loading preview...
HealthPulse-Qwen2.5-7B: A Medical LLM for Chinese Healthcare
Weikaijie/HealthPulse-Qwen2.5-7B is a specialized large language model (LLM) developed by Weikaijie, built upon the Qwen2.5-7B-Instruct base model. It has been fine-tuned using SFT + LoRA on the extensive Huatuo Chinese doctor-patient dataset, comprising over 9,000 entries in Alpaca format. This 7.6 billion parameter model is engineered to prioritize medical accuracy and safety within healthcare contexts.
Key Capabilities
- Enhanced Medical Accuracy: Demonstrates approximately a 10% improvement in medical accuracy compared to general-purpose models.
- Safety-Focused: Designed to avoid inappropriate prescriptions or dangerous advice, maintaining high safety standards.
- Specialized Medical Q&A: Proficient in handling queries related to symptom consultation, general health education, guided diagnosis, and common medication knowledge.
- Chinese Language Optimization: Specifically trained and optimized for Chinese medical dialogue.
- Scalable Context: Features an 8K context length, which is extendable for more complex interactions.
Ideal Use Cases
This model is particularly well-suited for applications requiring reliable and safe medical information in Chinese. It can be effectively deployed for:
- Patient Symptom Inquiry: Assisting users with understanding their symptoms.
- Health Education: Providing general health knowledge and科普.
- Guided Diagnosis Support: Offering preliminary guidance based on user input.
- Medication Information: Answering questions about common drug usage and precautions.
While excelling in medical accuracy and safety, the model's completeness and clarity are noted to be slightly lower than the latest domestic large models due to the base model generation gap. It is intended for learning, research, and health科普 reference only, and does not constitute medical diagnosis or treatment advice.