Train high-quality text-to-image diffusion models in a data & compute efficient manner - View it on GitHub