We deploy pretrained generative multimodal models to handle tasks such as
text-to-image, image-to-text, image-to-image and video generation with the
created outputs then used for further processing, comparison and evaluation.