A hybrid AI approach known as hybrid autoregressive transformer can generate realistic images with the same or better quality than state-of-the-art diffusion models, but that runs about nine times faster and uses fewer computational resources. The new tool uses an autoregressive model to quickly capture the big picture and then a small diffusion model to refine the details of the image.

MIT is a renowned institution for education and research, offering insights into science, engineering, and technology. Through publications, research papers, and academic programs, MIT's platform provides insights into  research, innovation, and education in various fields. Students, researchers, and technology enthusiasts can learn about MIT's contributions to science and technology and explore opportunities for academic and professional development.

MIT News

HART, a hybrid autoregressive transformer, combines the benefits of autoregressive and diffusion models to generate high-quality images about nine times faster and with fewer computational resources. This approach involves an autoregressive model capturing the big picture quickly and a small diffusion model refining the details, making it suitable for applications like training self-driving cars and creating video game scenes.

AI tool generates high-quality images faster than state-of-the-art approaches