RAG Document Q&A with Groq and Llama3

DEMO : https://akshaya-rag-document.streamlit.app/

This project is a Streamlit-based Q&A application that utilizes Retrieval-Augmented Generation (RAG) to answer questions from a collection of research papers. It combines Groq's Llama3 model for language understanding with FAISS for efficient vector search.

Features

• Document Ingestion: Automatically loads and processes research papers in PDF format.
• Embeddings Creation: Uses Ollama embeddings to create vector representations of document chunks.
• Vector Search: Uses FAISS to perform similarity search and retrieve relevant document chunks based on user queries.
• Q&A with Groq's Llama3: Leverages the power of Groq's Llama3 model to generate answers based on retrieved context.
• Streamlit UI: Simple and interactive user interface for entering queries and viewing results.

Prerequisites

• Python 3.8+

• A Groq API key for using the Llama3 model.

• Libraries: streamlit, langchain, langchain_community, faiss, dotenv

• Install all required libraries using the requirements.txt file:

  pip install -r requirements.txt

Acknowledgments

• LangChain for document processing and chaining.
• FAISS for vector search.
• Streamlit for creating a user-friendly interface.
• Groq API for their advanced language models.