Today, we’re showcasing CM3leon (pronounced like “chameleon”), a single foundation model that does both text-to-image and image-to-text generation.