What does it do?

1. Loads MNIST dataset
2. Builds two neural networks:
   • Generator: Transforms random noise into fake digit images
   • Discriminator: Distinguishes between real and fake images
3. Trains adversarially - Generator tries to fool discriminator, discriminator tries to detect fakes
4. Generates synthetic digits - Creates new handwritten digit images
5. Visualizes progress - Saves images at each epoch and creates animated GIF