ResearchBenchmarksVisual Grounding on RefCLEFFollow75.8AccuracyGiT-H72.05673.0287474.972Mar 14, 2024Evaluation ResultsMethodMethodLinksAccuracyGiT-HBackbone=Huge, Trainin...Backbone=Huge, Training=Universal2024.0375.8GiT-LBackbone=Large, Traini...Backbone=Large, Training=Universal2024.0374.5GiT-BBackbone=Base, Trainin...Backbone=Base, Training=Universal2024.0372.2