RVL-BERT: Visual Relationship Detection with Visual-Linguistic Knowledge from Pre-trained Representations

Open in new window