Once-Tuning-Multiple-Variants: Tuning Once and Expanded as Multiple Vision-Language Model Variants

Open paper

Benchmark	Model	Rank	Results
cross-modal-retrieval-on-flickr30k	BEiT-3 (Large)-OTMV	–	Image-to-text R@1: 97.3
visual-reasoning-on-nlvr2-test	BEiT-3 (Large)-OTMV	–	Accuracy: 89.96