CoolFace
Apppublic

Waris01/Multi-Model-Pdf-Chat

sourceHugging Faceapache-2.0updated 1y agoView on Hugging Face
0likes
App README

📄 Multi-Modal RAG PDF Chatbot

A Streamlit application that allows you to upload a PDF, ask questions about its content, and get accurate responses using a Multi-Modal Retrieval-Augmented Generation (RAG) pipeline powered by Groq's Gemma-2 9B model.

🚀 Features

  • —📁 Upload any PDF
  • —🔍 Intelligent chunking and embedding
  • —🧠 Ask natural language questions about your PDF
  • —⚡ Powered by FAISS + HuggingFace + Groq LLM
  • —🧠 Caches session so PDF isn't reprocessed on every query

🛠️ Installation (with venv)

  1. 1.Clone the repo:
bash
git clone https://github.com/Warishayat/Multimodel-Rag-Application01.git
cd Multimodal-Rag-Application01
  1. 1.Create and activate a virtual environment:
bash
python -m venv venv
# Activate:
# On Windows
venv\Scripts\activate
# On macOS/Linux
source venv/bin/activate
  1. 1.Install dependencies:
bash
pip install -r requirements.txt
  1. 1.Set up your `.env` file:

Create a .env file in the root directory:

GROQ_API_KEY=your_groq_api_key_here

📦 Project Structure

📁 Multimodal-Rag-Application01
├── main.py                 # Streamlit frontend
├── pdfparsing.py          # PDF parser using pymupdf4llm
├── Datapreprocessing.py   # Chunking & text cleaning
├── vectorstore.py         # Embedding & FAISS logic
├── .env                   # API keys
├── requirements.txt       # Python dependencies
└── README.md              # You're here!

▶️ Run the App

bash
streamlit run main.py

Then open http://localhost:8501 in your browser.

🧪 Example Queries

After uploading a PDF, try asking:

  • —"What is the summary of section 3?"
  • —"List all benchmarks mentioned."
  • —"How is this model different from others?"

💡 Tips

  • —PDF is processed only once per session using st.session_state.
  • —Uses RecursiveCharacterTextSplitter for effective chunking.
  • —Embedding with HuggingFaceEmbeddings.

📋 Requirements

Make sure your requirements.txt includes at least:

txt
streamlit
python-dotenv
langchain
langchain-community
langchain-groq
faiss-cpu
pymupdf4llm

📬 Credits

Built with ❤️ by Waris Hayat Abbasi.