--- task_categories: - image-to-text - text-to-image - visual-question-answering language: - en - vi size_categories: - 10M