🦙 Llama3 RAG Wiki

Retrieval-Augmented Generation (RAG) with Llama 3 and Wikipedia | Local Open-Source LLM Chatbot

---

🔍 What is Llama3 RAG Wiki?

Llama3 RAG Wiki is a local, open-source Retrieval-Augmented Generation (RAG) chatbot built using Llama 3, Ollama, and Wikipedia.

It demonstrates how to:

• Combine LLMs + semantic search
• Reduce hallucinations using external knowledge retrieval
• Build a fully local RAG pipeline
• Implement bare-bones-style RAG in Python

This project is ideal for LLM engineers, AI researchers, students, and open-source contributors looking to understand or build RAG systems from scratch.

---

✨ Key Features

• 🧠 Local Llama 3 (8B) inference via Ollama
• 📚 Real-time Wikipedia-based knowledge retrieval
• 🔍 Semantic search using Sentence Transformers
• 🧩 Modular RAG architecture
• 📓 Step-by-step Jupyter Notebook tutorial
• 🖥️ Standalone Python CLI application
• 🔓 100% open-source and offline-friendly

---

🏗️ RAG Architecture Overview

This project follows a standard Retrieval-Augmented Generation pipeline:

1. User submits a query
2. Relevant Wikipedia articles are retrieved
3. Text is chunked and embedded
4. Semantic similarity search selects top context
5. Context is injected into the LLM prompt
6. Llama 3 generates a grounded response

Architecture Diagram

graph LR
    A[User Query] --> B[Wikipedia API]
    B --> C[Wikipedia Articles]
    C --> D[Text Chunking]
    D --> E[Embedding Model<br/>gte-base-en-v1.5]
    E --> F[Vector Similarity Search]
    F --> G[Top-K Relevant Chunks]
    G --> H[Prompt Augmentation]
    H --> I[Llama 3 LLM<br/>via Ollama]
    I --> J[Final Answer]

Loading

---

📘 Learn More

📖 LinkedIn Article: A beginner-friendly explanation of LLMs and RAG architecture:

👉 Explain LLM + RAG Like I’m 5

---

🔄 Project Variants

The repository includes two implementations:

📓 Jupyter Notebook

• Step-by-step explanation of RAG internals
• Ideal for learning and experimentation

🖥️ Python Application

• End-to-end local RAG chatbot
• Suitable for real-world usage and demos

---

🧠 Models Used

Component	Model
LLM	Llama 3 (8B)
Embeddings	Alibaba-NLP/gte-base-en-v1.5

---

📦 Dependencies

• ollama – v0.2.1
• sentence-transformers – v3.0.1
• numpy – v1.26.4
• Wikipedia-API – v0.6.0

---

⚙️ Installation & Setup

Prerequisites

• Python 3.9+
• Ollama installed locally

Pull Required Models

ollama pull llama3
ollama pull llama3.1

Run the Application

python Llama3_RAG_Wiki.py

---

🎯 Use Cases

• 🧪 Learning Retrieval-Augmented Generation
• 🤖 Building local AI chatbots
• 📚 Question answering over external knowledge
• 🛠️ LLM system design experimentation
• 🎓 AI education & workshops

---

🌱 Future Enhancements

• Vector database integration (FAISS / Chroma)
• Multi-document retrieval
• Query rewriting and reranking
• Streaming responses
• Web-based UI

---

⭐ Why This Repo Matters

• Demonstrates real-world RAG implementation
• Uses state-of-the-art open-source LLMs
• Runs entirely on your local machine
• Beginner-friendly yet production-aligned

If this project helped you, please consider giving it a ⭐!

---