medical-assistant-muxi

Medical Science Assistant

Project Overview

This project is an intelligent medical Q&A assistant based on Moore Threads GPU. Leveraging domestic Moore Threads GPU computing power, it deploys the Qwen2.5-0.5B large language model to provide users with medical science knowledge including symptom interpretation, health advice, and preventive measures for common diseases. The project aims to explore applications of domestic computing platforms in healthcare and promote AI technology for public health. ##Core Features *Common Disease Q&A:Supports consultations for colds, hypertension, diabetes, fever, etc. *Symptom Interpretation:Helps users understand disease symptoms and manifestations *Health Recommendations:Provides scientific and professional prevention/treatment advice *User-Friendly: Command-line interaction with instant responses ##Technology Stack Component Technology LLM Qwen2.5-0.5B Computing Platform Moore Threads GPU (Xiyun C500) Inference Framework Transformers + PyTorch Runtime Environment Gitee AI Model Ark ##Usage

#enter project directory
cd /data/medical_assistant
#Run
python medical_assistant.py

Running results.

Loading the model Model loading completed

---

the medical assistant is start enter your problems

---

your problems:How to prevent the common cold? the users’ problems:How to prevent the common cold? please answer with simple words: To prevent the common cold, you can wash your hands often, avoid close contact with infected individuals, and maintain a balanced diet and exercise routine. You can also drink plenty of water and take care of your skin and hair to prevent dryness and irritation. your problems:quit #Project Value Insights 1.Domestic Tech Validation: Demonstrates successful deployment of China’s Moore Threads GPU for medical AI applications, proving its capability in handling LLM inference tasks. 2.Public Health Impact: Lowers medical knowledge barriers through 24/7 accessible symptom interpretation and prevention advice (e.g., hypertension/diabetes management). 3.Open Innovation: MIT-licensed codebase enables hospitals/researchers to customize modules for specific diseases or integrate with EHR systems. Key extensibility includes: ·Multilingual support adaptation ·Regional disease database integration ·GPU optimization techniques sharing #References ·Qwen2.5 - Alibaba Cloud’s Tongyi Qianwen open-source model ·Moore Threads GPU - China’s high-performance GPU ·Gitee AI - Model Ark computing platform #Demo video. click here to watch the demo vedio:medical_assistant_demo.mp4