Built using TensorFlow.js for running the GAN model in the browser, p5.js for canvas manipulation and visualization, and vanilla JavaScript for interaction handling. The model was pre-trained on a dataset of abstract art and optimized for real-time inference.

The biggest challenge was optimizing the GAN model to run smoothly in the browser without compromising the quality of generated art. I learned about model quantization, web workers for background processing, and efficient canvas rendering techniques. This project deepened my understanding of how to bridge the gap between complex ML models and user-friendly web interfaces.