DTOS: Dynamic Time Object Sensing with Large Multimodal Model

Benchmark Model Rank Results
referring-video-object-segmentation-on-refvos-davisDTOS-9B#2J&F: 74.88
moment-retrieval-on-charades-staDTOS-TCS#13R@1 IoU=0.5: 69.81R@1 IoU=0.7: 43.2R@1 IoU=0.3: 81.29mIoU: 58.0
referring-video-object-segmentation-on-mevisDTOS-9B#19J&F: 48.86
moment-retrieval-on-qvhighlightsDTOS-TCS#27mAP: 40.98R@1 IoU=0.5: 72.24R@1 IoU=0.7: 49.22mAP@0.5: 64.75…