Fetching the paper…

ROME: Evaluating Pre-trained Vision-Language Models on Reasoning beyond Visual Common Sense · Around