Latent-conditioned Denoising Diffusion-based Probabilistic Model (DDPM)

Final project for fundamentals of deep learning at Toyota Tech. Institute Chicago with Professor David A. McAllester. The goal of the final project is to extend the DDPM work of Ho et al., 2020 such that we infer low-dimensional latent representations that is used by the Unet model in the DDPM.