Generates random shapes including dots, lines, polygons, circles, and stars and trains a U-Net to fill them in. Uses Pillow for image generation and PyTorch for designing the U-Net architecture.

To get the code:

git pull https://github.com/Carson274/Shape-Filler.git

1. Test the currently loaded model:

python3 test.py

2. Train a new model:

python3 train.py

• PyTorch: Neural network modeling and training.
• Pillow (PIL): Image generation and processing.
• NumPy: Data manipulation and operations.
• torchvision: Image transformations and utilities.
• matplotlib: Plotting training loss graphs.