3 ms·Unified Visual Relationship Detection with Vision and Language Models1 points by georgehill 4y ago