Session
From U-Nets to Diffusion: Build a U-Net, a type of autoencoder for images; Learn about transposed convolution to increase the size of an image; Learn about non-sequential neural networks and residual connections; Experiment with feeding noise through the U-Net to generate new images