🚀 Gemma Model Fine-tuning UI

A comprehensive and user-friendly web interface for fine-tuning Google's Gemma models on custom datasets without requiring deep ML expertise.

🌟 Features

• 📂 Dataset Management

  • Support for CSV, JSONL, and text files
  • Automated validation and preprocessing
  • Data augmentation options
  • Dataset preview and statistics

• 🎛️ Hyperparameter Configuration

  • Intuitive UI for parameter adjustment
  • Sensible defaults with explanations
  • Configuration templates for common use cases
  • Parameter validation

• 📊 Training Visualization

  • Real-time loss curves
  • Evaluation metrics tracking
  • Resource utilization monitoring
  • Example generation during training

• 💾 Model Export Options

  • Download fine-tuned models in various formats (PyTorch, TensorFlow, GGUF)
  • Direct Hugging Face Hub integration
  • Model compression options
  • Deployment configuration generation

• ☁️ Cloud Integration (Coming Soon)

  • Google Cloud Storage support
  • Vertex AI training capabilities
  • Distributed training configuration
  • TPU acceleration options

📋 Prerequisites

• Python 3.8 or higher
• Access to Gemma models (requires Google AI Studio access)
• For TPU support: GCP account with TPU access

🔧 Installation

1. Clone the repository:

   git clone https://github.com/Taskmaster-1/Gemma-Model-Fine-tuning-UI.git
   cd Gemma-Model-Fine-tuning-UI

2. Create a virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies:

   pip install -r requirements.txt

4. Set up Gemma model access:

   • Generate an API key from Google AI Studio
   • Configure your environment:

     export GOOGLE_API_KEY="your_api_key_here"
     # On Windows: set GOOGLE_API_KEY=your_api_key_here

🚀 Quick Start

1. Launch the application:

   streamlit run ui/app.py

2. Open your browser and navigate to http://localhost:8501

3. Upload your dataset, select a model, configure parameters, and start fine-tuning!

📚 Documentation

For detailed documentation, see the docs directory:

• User Guide - Complete usage instructions
• Developer Guide - How to contribute and extend
• API Reference - API documentation
• Examples - Example use cases and workflows

🤝 Contributing

Contributions are welcome! This project is part of Google Summer of Code. See CONTRIBUTING.md for details on how to contribute.

📜 License

This project is licensed under the MIT License - see the LICENSE file for details.

🔗 Related Projects

• Gemma - Official Gemma repository
• HuggingFace Transformers - Used for model implementation

📞 Contact

For questions or support, please open an issue or contact the author at i.am.vivekyadav5223@gmail.com.

👏 Acknowledgements

• Google for creating and open-sourcing the Gemma models
• The Hugging Face team for their transformers library
• All contributors to this project