LLaMa3-RAG Document Intelligence System

Website URL: https://satyadewangan05-rag-using-llama3-web-jeqers.streamlit.app/

This project implements a Retrieval-Augmented Generation (RAG) system using LLaMa3 and LlamaIndex, deployed with FastAPI and Streamlit. It enables efficient, user-friendly access to advanced AI-powered document retrieval and generation capabilities.

Features

• Utilizes LLaMa3 for advanced natural language processing
• Implements efficient document indexing and semantic search using LlamaIndex
• FastAPI backend for robust API endpoints
• Streamlit frontend for an intuitive user interface
• Supports document upload and intelligent querying

How to Use the App

Follow these steps to set up and use the application:

1. Run the Colab Notebook

   • 

   • Open and run the provided Google Colab notebook

   • This will start the FastAPI backend and set up ngrok for public access

2. Get the ngrok URL

   • Look for the ngrok URL in the Colab notebook output
   • It should look something like https://xxxx-xx-xx-xxx-xx.ngrok.io

3. Access the Streamlit Web Interface

   • Open the Streamlit app in your web browser
   • Paste the ngrok URL into the designated input field on the sidebar

4. Upload a Document

   • Use the file upload feature in the Streamlit interface
   • Support file types include .txt, .pdf, and .docx (adjust as necessary for your implementation)

5. Start Chatting

   • Once your document is uploaded, you can start asking questions
   • The system will use the uploaded document as context for answering your queries

Technical Details

• Backend: FastAPI
• Frontend: Streamlit
• AI Model: LLaMa3
• Vector Store: LlamaIndex
• Deployment: Google Colab with ngrok for public access

Notes

• Ensure you have a stable internet connection while using the app
• The ngrok URL will change each time you run the Colab notebook
• For persistent deployment, consider using a cloud platform instead of Colab and ngrok

Future Improvements

• Add support for multiple document uploads
• Implement user authentication for personalized experiences
• Enhance the UI with more interactive visualizations of the document retrieval process

For any issues or suggestions, please open an issue in this repository.