🤖 What You Can Build with Gemma 3n Now

This repository contains practical applications and examples demonstrating the capabilities of Google's Gemma 3n, the latest multimodal AI model optimized for on-device inference.

📁 Created Applications

gemma3n_multimodal_chat.py

Type: Streamlit Web App
Description: Multimodal chat application supporting text, image, and audio inputs
Features: Multiple API providers, Real-time chat, File uploads, Conversation history
Complexity: Intermediate

gemma3n_voice_assistant.py

Type: Streamlit Voice App
Description: Voice-activated AI assistant with speech recognition and text-to-speech
Features: Voice input/output, Multiple languages, Conversation context, Manual text fallback
Complexity: Advanced

gemma3n_coding_agent.py

Type: Command Line Tool
Description: Local coding assistant for code analysis, generation, and debugging
Features: Interactive chat, Code analysis, Project generation, Debug assistance
Complexity: Advanced

gemma3n_document_analyzer.py

Type: Streamlit Web App
Description: Document analysis tool for PDFs and images with AI-powered insights
Features: PDF processing, Multiple analysis types, Data export, Batch processing
Complexity: Advanced

requirements.txt

Type: Configuration
Description: Python package dependencies for all applications
Features: Complete dependency list, Version specifications, Optional packages
Complexity: Basic

setup.sh

Type: Installation Script
Description: Automated setup script for Ollama and Python environment
Features: Cross-platform support, Ollama installation, Virtual environment setup
Complexity: Intermediate

docker-compose.yml

Type: Docker Configuration
Description: Docker Compose configuration for containerized deployment
Features: Ollama service, App containerization, Network configuration
Complexity: Intermediate

Dockerfile

Type: Docker Configuration
Description: Docker container definition for the applications
Features: Python environment, System dependencies, Non-root user
Complexity: Intermediate

🚀 Quick Start

Prerequisites

• Python 3.8+
• Ollama installed and running
• Gemma 3n model downloaded

Installation

1. Automated Setup (Recommended):

   chmod +x setup.sh
   ./setup.sh

2. Manual Setup:

   # Install Ollama
   curl -fsSL https://ollama.com/install.sh | sh
   
   # Download Gemma 3n
   ollama pull gemma3n:e4b
   ollama pull gemma3n:e2b
   
   # Install Python dependencies
   pip install -r requirements.txt

Running Applications

1. Multimodal Chat:

   streamlit run gemma3n_multimodal_chat.py

2. Voice Assistant:

   streamlit run gemma3n_voice_assistant.py

3. Coding Agent:

   python gemma3n_coding_agent.py --interactive

4. Document Analyzer:

   streamlit run gemma3n_document_analyzer.py

🐳 Docker Deployment

For containerized deployment:

docker-compose up -d

💡 Use Cases Demonstrated

• Multimodal Chat: Customer support, education, content creation
• Voice Assistant: Accessibility, hands-free operation, smart home integration
• Coding Agent: Development assistance, code review, debugging
• Document Analysis: Legal document review, research, data extraction

🔧 Customization

Each application is designed to be easily customizable:

• API Providers: Switch between Ollama, Together AI, Google AI Studio
• Models: Support for both E2B and E4B variants
• UI/UX: Streamlit components can be modified or replaced
• Features: Add new analysis types, commands, or integrations

📚 Learning Resources

• Gemma 3n Documentation
• Ollama Documentation
• Streamlit Documentation

🤝 Contributing

Feel free to submit issues, feature requests, or pull requests to improve these applications.

📄 License

This project is open source and available under the MIT License.