Generative AI Course 🤖

A comprehensive collection of Jupyter notebooks covering the essential aspects of Generative AI, from foundational concepts to advanced applications. This repository serves as a practical guide for learning and implementing modern AI techniques.

🎯 Overview

This course provides hands-on experience with:

• Retrieval-Augmented Generation (RAG) systems
• DSPy framework for programming with language models
• Model fine-tuning with Google's Gemma and OpenAI models
• Quantization techniques for model optimization
• Agentic AI and multi-agent systems
• Transformers library exploration
• GGUF model conversion and deployment

📚 Course Content

🔍 RAG (Retrieval-Augmented Generation)

• simple_rag.ipynb - Basic RAG implementation using LangChain and ChromaDB
• rag_with_dspy.ipynb - Advanced RAG with DSPy framework integration
• agentic_rag.ipynb - Agentic RAG using LangGraph and NVIDIA AI Endpoints

🧠 DSPy Framework

• dspy_demo.ipynb - Introduction to DSPy for prompt optimization and language model programming

🎛️ Model Fine-tuning

• Finetune_Gemma.ipynb - Complete guide to fine-tuning Google's Gemma model
• gpt-oss-ft.ipynb - Fine-tuning OpenAI's GPT models (open-source)

⚡ Model Optimization

• intro_to_quant.ipynb - Introduction to quantization techniques for model compression
• quantize_gemma_gguf.ipynb - Converting and quantizing Gemma models to GGUF format

🤝 Multi-Agent Systems

• arxiv_research_crew.ipynb - Building research crews using CrewAI for arXiv paper analysis

🔧 Framework Exploration

• Tour_of_Transformers_Library.ipynb - Comprehensive tour of HuggingFace Transformers library

🚀 Quick Start

Prerequisites

# Create a virtual environment
python -m venv genai-course
source genai-course/bin/activate  # Linux/Mac
# or
genai-course\Scripts\activate  # Windows

# Install core dependencies
pip install jupyter numpy pandas matplotlib

For RAG Tutorials

pip install langchain langchain-community chromadb gpt4all
pip install beautifulsoup4 tqdm

For DSPy Tutorials

pip install dspy-ai torch transformers accelerate

For Fine-tuning Tutorials

pip install transformers accelerate datasets peft
pip install torch torchvision torchaudio

For Quantization

pip install transformers accelerate matplotlib numpy

For Multi-Agent Systems

pip install crewai crewai_tools arxiv

📁 Repository Structure

├── README.md                           # This file
├── simple_rag.ipynb                   # Basic RAG implementation
├── rag_with_dspy.ipynb                # Advanced RAG with DSPy
├── agentic_rag.ipynb                  # Agentic RAG systems
├── dspy_demo.ipynb                    # DSPy framework demo
├── Finetune_Gemma.ipynb               # Gemma fine-tuning guide
├── gpt-oss-ft.ipynb                   # GPT fine-tuning
├── intro_to_quant.ipynb               # Quantization introduction
├── quantize_gemma_gguf.ipynb          # GGUF conversion & quantization
├── arxiv_research_crew.ipynb          # Multi-agent research system
├── Tour_of_Transformers_Library.ipynb # Transformers library tour
├── llama.cpp/                         # GGUF conversion tools
├── merged_model/                      # Merged model outputs
├── peft_model/                        # PEFT adapter models
└── *.log                             # Training and usage logs

🎯 Learning Path

Beginner Track

1. Tour_of_Transformers_Library.ipynb - Get familiar with the ecosystem
2. simple_rag.ipynb - Understand RAG fundamentals
3. intro_to_quant.ipynb - Learn model optimization basics

Intermediate Track

4. dspy_demo.ipynb - Advanced prompt engineering
5. rag_with_dspy.ipynb - Sophisticated RAG systems
6. Finetune_Gemma.ipynb - Custom model training

Advanced Track

7. agentic_rag.ipynb - Multi-agent AI systems
8. arxiv_research_crew.ipynb - Complex workflow automation
9. quantize_gemma_gguf.ipynb - Production deployment

🛠️ Key Technologies

• 🤗 HuggingFace Transformers - Model hub and training
• 🦜🔗 LangChain - LLM application framework
• 📊 ChromaDB - Vector database for embeddings
• 🧪 DSPy - Programming with language models
• ⚡ PEFT - Parameter-efficient fine-tuning
• 🔧 llama.cpp - Efficient model inference
• 👥 CrewAI - Multi-agent orchestration
• 🎯 NVIDIA AI Endpoints - GPU-accelerated inference

📊 Models Covered

• Google Gemma (2B parameters) - Fine-tuning and quantization
• OpenAI GPT models - Fine-tuning workflows
• Various HuggingFace models - For RAG and embeddings
• Custom fine-tuned models - PEFT adapters and merging

🎓 What You'll Learn

• RAG Systems: Build intelligent document retrieval and question-answering systems
• Model Fine-tuning: Customize large language models for specific tasks
• Quantization: Optimize models for deployment and inference speed
• DSPy Programming: Advanced prompt engineering and model chaining
• Multi-Agent AI: Create collaborative AI systems for complex tasks
• Production Deployment: Convert and optimize models for real-world use

🔧 Hardware Requirements

• Minimum: 8GB RAM, CPU-only inference possible
• Recommended: 16GB+ RAM, NVIDIA GPU with 8GB+ VRAM
• For Fine-tuning: 24GB+ VRAM recommended (or use gradient checkpointing)

📝 Usage Notes

• Each notebook is self-contained with installation instructions
• Some notebooks require API keys (OpenAI, HuggingFace, NVIDIA)
• Models are cached locally after first download
• Large model files are not included in the repository

🤝 Contributing

Contributions are welcome! Please:

1. Fork the repository
2. Create a feature branch
3. Add comprehensive documentation
4. Submit a pull request

📄 License

This project is open source. Please check individual model licenses when using specific models.

🙏 Acknowledgments

• Google Research for Gemma models
• HuggingFace for the Transformers library
• LangChain community for RAG frameworks
• DSPy team for advanced LLM programming
• llama.cpp project for efficient inference

---

Happy Learning! 🚀

For questions or issues, please open a GitHub issue or reach out to the maintainers.