ResearchBenchmarksQuery-based video segmentation (Text→Audio) on AVSBench S4 (VGGSound-AVEL 40K)Follow77.7mIoUDCID75.30875.92976.5577.171May 12, 2026Evaluation ResultsMethodMethodLinksmIoUF-scoreDCIDSetting=AVT, Fine-tune...Setting=AVT, Fine-tuned epoch=4th2026.0577.786.7MICUSetting=AVT, Fine-tune...Setting=AVT, Fine-tuned epoch=4th2026.0576.986.5CoDAARSetting=AVT, Fine-tune...Setting=AVT, Fine-tuned epoch=4th2026.0575.486.6