Researchers from Caltech and ETH Zurich Introduce Groundbreaking Diffusion Models: Harnessing Text Captions for State-of-the-Art Visual Tasks and Cross-Domain Adaptations