Skip to content

Latest commit

 

History

History
63 lines (52 loc) · 2.16 KB

README.md

File metadata and controls

63 lines (52 loc) · 2.16 KB

MedQA: Medical Question-Answering System

Python LangChain Llama License

Table of Contents

Overview

MedQA is an advanced medical question-answering system that leverages the power of Llama 3.1-8B model and Retrieval-Augmented Generation (RAG) to provide accurate and contextual answers to medical queries. Built on the PubMedQA dataset, this system combines state-of-the-art language models with efficient information retrieval techniques.

Features

  • Accurate Medical QA: Fine-tuned for medical domain questions
  • RAG-Enhanced Responses: Combines retrieval and generation for evidence-based answers
  • User-Friendly Interface: Gradio-based web application
  • Source Transparency: Provides references for generated answers
  • Efficient Processing: Optimized with LoRA fine-tuning
  • Scalable Architecture: Modular design for easy updates and improvements

Architecture

System Architecture

Components:

  1. Model Layer

    • Llama 3.1-8B base model
    • LoRA fine-tuning adaptations
    • Custom stop sequences and sampling controls

  2. Retrieval Layer

    • FAISS vector store
    • Sentence-transformer embeddings
    • Top-k document retrieval

  3. Integration Layer

    • LangChain RetrievalQA chain
    • Custom LoRAMedicalLLM implementation
    • Gradio web interface

Installation

# Clone the repository
git clone https://github.com/VijayendraDwari/MedQA.git
cd MedQA

# Create virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt