https://github.com/shubhammandowara/llm_rag
🚀 Transform Any PDF into an AI-Powered Q&A Chatbot!
https://github.com/shubhammandowara/llm_rag
faiss huggingface langchain llm llm-rag llm-rag-chatbot pdf-chat-bot pdf-chatbot qna questions-and-answers rag streamlit vector-database
Last synced: about 2 months ago
JSON representation
🚀 Transform Any PDF into an AI-Powered Q&A Chatbot!
- Host: GitHub
- URL: https://github.com/shubhammandowara/llm_rag
- Owner: ShubhamMandowara
- License: gpl-3.0
- Created: 2025-02-04T13:53:19.000Z (4 months ago)
- Default Branch: main
- Last Pushed: 2025-02-28T12:50:50.000Z (3 months ago)
- Last Synced: 2025-02-28T18:48:08.302Z (3 months ago)
- Topics: faiss, huggingface, langchain, llm, llm-rag, llm-rag-chatbot, pdf-chat-bot, pdf-chatbot, qna, questions-and-answers, rag, streamlit, vector-database
- Language: Python
- Homepage: https://pdfllmrag.streamlit.app/
- Size: 67.4 KB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# 📖 PDF Q&A Chatbot using Streamlit & Hugging Face API
## 🚀 Overview
This project is a **PDF-based Q&A chatbot** built with **Streamlit**, **FAISS** for similarity search, and **Hugging Face Inference API** for generating answers. Users can upload a **PDF file**, ask questions about its content, and receive answers using an **LLM (e.g., Mistral-7B, Falcon-7B, Llama-3, etc.)**.
---
## 🎯 Features
✅ **Upload PDFs** & extract text + tables 📄
✅ **Vector search with FAISS** for retrieving relevant text 🔍
✅ **Generate answers using Hugging Face Inference API** 🤖
✅ **Streamlit UI for easy interaction** 🖥️
✅ **Supports any Hugging Face text generation model** 🔄
---
## 🛠️ Installation
### **1️⃣ Clone the Repository**
```bash
git clone https://github.com/ShubhamMandowara/llm_rag.git
cd pdf-qna-chatbot
```
### **2️⃣ Create a Virtual Environment**
```bash
python -m venv venv
source venv/bin/activate # On Windows: venv\Scripts\activate
```
### **3️⃣ Install Dependencies**
```bash
pip install -r requirements.txt
```
---
## 🔑 Set Up Hugging Face API Token
1. Get your API token from [Hugging Face Tokens](https://huggingface.co/settings/tokens).
2. When running the Streamlit app, enter the API token in the provided input field.
3. Or, store it in **Streamlit secrets** by adding to `.streamlit/secrets.toml`:
```toml
[secrets]
HUGGINGFACEHUB_API_TOKEN = "your_token_here"
```
---
## 🎯 Usage
### **Run the Streamlit App**
```bash
streamlit run app.py
```
### **How It Works**
1️⃣ **Upload a PDF** to extract text & tables.
2️⃣ **FAISS retrieves relevant text chunks** from the document.
3️⃣ **Hugging Face Inference API** generates an answer based on context.
4️⃣ **Streamlit UI displays the answer.** 🎉
---
## 🏗️ Project Structure
```
📂 pdf-qna-chatbot
│── app.py # Main Streamlit app
│── requirements.txt # Required dependencies
│── secrets.toml # Streamlit config folder (for secrets)
```
---
## 📌 Example Models (Change in `app.py`)
- **Mistral-7B**: `mistralai/Mistral-7B-Instruct`
- **Llama-3**: `meta-llama/Llama-3-8B`
- **Falcon-7B**: `tiiuae/falcon-7b-instruct`
To change the model, update:
```python
HF_MODEL = "mistralai/Mistral-7B-Instruct" # Change to your preferred model
```
---
## 🛠️ Troubleshooting
### **1️⃣ Model Not Found / API Issues**
- Ensure your **Hugging Face API token** is correct.
- Check if the model is **public** and supports the **Inference API**.
- Try testing API manually:
```bash
curl -H "Authorization: Bearer your_api_token" https://api-inference.huggingface.co/models/mistralai/Mistral-7B-Instruct
```
### **2️⃣ FAISS Index Not Working**
- Ensure FAISS is installed:
```bash
pip install faiss-cpu # Or faiss-gpu if using CUDA
```
### **3️⃣ Streamlit Not Running**
- Restart the app after setting environment variables:
```bash
streamlit run app.py
```
---
## 📜 License
This project is **open-source** under the GPL-3.0 License.
---
## 👨💻 Author
Developed by [Shubham Mandowara](https://github.com/shubhammandowara). Contributions are welcome! 😊🚀