This repository implements a face depixelizer based on the PULSE (Self-Supervised Photo Upsampling via Latent Space Exploration of Generative Models) paper. Given a low-resolution face image, it searches the latent space of StyleGAN to find high-resolution outputs that downscale correctly to match the input. The system uses a generative model to produce perceptually realistic face images from pixelated inputs, providing Jupyter notebooks for running inference on Google Colab.