This PR updates the dataset card by adding `text-to-image` to the `task_categories` metadata. This addition accurately reflects that the dataset contains a large number of images generated by various text-to-image models (among others), making it highly relevant for research and benchmarking in the text-to-image domain.
