This is the easiest way to run multiple image models in parallel, see their outputs side-by-side, and glance at their runtimes.