ResearchBenchmarksQuery-based video segmentation (Text→Audio) on AVSBench S4 VGGSound-AVEL 90KFollow76.2mIoUDCID75.6875.81575.9576.085May 12, 2026Evaluation ResultsMethodMethodLinksmIoUF1 ScoreDCIDSetting=AVT, Fine-tune...Setting=AVT, Fine-tuned epoch=4th2026.0576.286.3CoDAARSetting=AVT, Fine-tune...Setting=AVT, Fine-tuned epoch=4th2026.057686MICUSetting=AVT, Fine-tune...Setting=AVT, Fine-tuned epoch=4th2026.0575.785.7