DeePTB Recipes 📚

This repository contains tutorials and examples for DeePTB - a deep learning package for accelerating ab initio electronic structure simulations.

🚀 Quick Start

All tutorials are available in two versions:

• Standard Jupyter Notebooks: For local execution
• Google Colab Notebooks: For online execution (no installation required!)

📖 DeePTB Tutorials (v2.2)

Tutorial 1: DeePTB-SK Baseline Model

Learn how to use built-in base models to plot band structures for given crystal structures.

---

Tutorial 2: Training DeePTB-SK Model

Learn how to train a DeePTB-SK model from scratch using first-principles data.

---

Tutorial 2.1: Advanced DeePTB-SK Training

Advanced training techniques and optimization strategies for DeePTB-SK models.

---

Tutorial 3: DeePTB-E3 Model

Learn how to use E3-equivariant neural networks for representing quantum operators.

---

Tutorial 4: Advanced Applications

Advanced applications and use cases of DeePTB models.

---

💻 Running Locally

Prerequisites

• Python 3.10+
• UV package manager (recommended)

Installation

# Clone DeePTB repository
git clone https://github.com/deepmodeling/DeePTB.git
cd DeePTB

# Install using UV
uv sync

# Or install using pip
pip install -e .

Run Tutorials

# Clone this repository
git clone https://github.com/DeePTB-Lab/Recipes.git
cd Recipes/deeptb_tutorials/v2.2

# Launch Jupyter
jupyter notebook

☁️ Running on Google Colab

Simply click the "Open in Colab" badge above any tutorial! The Colab version will:

• ✅ Automatically detect your environment (GPU/CPU)
• ✅ Install DeePTB and all dependencies
• ✅ Download required data files
• ✅ Ready to run in 5-7 minutes

💡 Tip: First-time setup takes 5-7 minutes. Subsequent runs will be faster if you keep the runtime alive.

📦 What's Included

• Tutorials: Step-by-step guides for using DeePTB
• Data: Example datasets for training and testing
• Scripts: Utility scripts for data processing and conversion

🔧 Technical Details

Colab Setup Features

• Automatic environment detection (Colab/Binder/Local)
• Smart CUDA version detection (nvidia-smi → torch → default)
• Fallback installation methods (UV → pip)
• Progress indicators and detailed logging
• Support for repeated execution

Installation Flow

1. Environment detection
2. UV package manager installation
3. DeePTB repository cloning
4. Dependency installation (PyTorch, torch_scatter, etc.)
5. Tutorial data download
6. Installation verification

📚 Documentation

• DeePTB Documentation
• DeePTB GitHub
• Installation Guide
• Troubleshooting

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

📄 License

This project is licensed under the LGPL-3.0 License - see the LICENSE file for details.

🙏 Acknowledgments

• DeePTB development team
• DeepModeling community
• All contributors to this repository

---

Author: Gu, Qiangqiang (顾强强)
Email: guqq@ustc.edu.cn
Date: 2025-11-21
Version: v2.2