CVA: Context-aware Video-text Alignment for Video Temporal Grounding

Benchmark Model Rank Results
moment-retrieval-on-charades-staCVA (extra data)–R@1 IoU=0.5: 62.61R@1 IoU=0.7: 40.78R@1 IoU=0.3: 74.19mIoU: 53.35
moment-retrieval-on-qvhighlightsCVA (extra data)–mAP: 47.49R@1 IoU=0.5: 70.05R@1 IoU=0.7: 55.32mAP@0.5: 69.49…
natural-language-moment-retrieval-on-tacosCVA (extra data)–R@1,IoU=0.3: 58.80R@1,IoU=0.5: 43.21R@1,IoU=0.7: 27.73mIoU: 41.07