A modern, high-performance desktop application built with Python, CustomTkinter, PyTorch, FastAI, and OpenCV to colorize and restore black-and-white photos using offline AI models.
If you just want to run the app on Windows without installing Python or dependencies:
- 📦 Download Portable Executable Version (v1.0.0 .ZIP)
- Extract the
.ziparchive on your computer. - Download the 5 required model weights from the table below and place them inside the
models/directory. - Launch
AI_Photo_Colorizer_Pro.exeand enjoy!
If you prefer running the app via Python source code:
- Clone or Download this repository to your local machine.
- Download the required model weights (listed below) and place them inside the
models/directory. - Double-click
run_app.bat. It will automatically set up the virtual environment, install dependencies, and launch the application!
- Multi-Model AI Engine: Switch seamlessly between Zhang (Caffe), DeOldify Artistic, and DeOldify Stable models.
- Modern Dark Interface: Built with CustomTkinter featuring clean dark themes, rounded widgets, and real-time live preview.
- Hardware Acceleration: Auto-detects NVIDIA CUDA GPU acceleration or smoothly falls back to CPU.
- Post-Processing Enhancements: Real-time controls for Saturation, Contrast, Brightness, Sharpness, and Gamma.
- Dynamic Controls: Toggle enhancements on/off with visual status indicators.
All required model files are hosted directly in the Releases section of this repository:
| Model File | Description | Download Link |
|---|---|---|
colorization_deploy_v2.prototxt |
Zhang Caffe Model Architecture | Download |
pts_in_hull.npy |
Zhang Caffe Points Data | Download |
colorization_release_v2.caffemodel |
Zhang Caffe Weights | Download |
ColorizeArtistic_gen.pth |
DeOldify Artistic Weights | Download |
ColorizeStable_gen.pth |
DeOldify Stable Weights | Download |
AI-Photo-Colorizer-Pro/
├── models/ <-- Place downloaded model weights here
├── dummy/ <-- Auto-created workspace folder
├── result_images/ <-- Auto-created output folder
├── app.py <-- Main CustomTkinter GUI application
├── run_app.bat <-- One-click Windows launcher script
├── requirements.txt <-- Python package dependencies list
└── README.md <-- Project documentation