ResearchBenchmarksVideo Semantic Segmentation on OVISVideo Semantic Segmentation on OVISFollow18.6mAPCLIP-VIS1.0245.58710.1514.713Jan 22, 2026Evaluation resultsMethodMethodLinksmAPmIoUCLIP-VISTraining=supervisedTraining=supervised2026.0118.660.09PyraTokTraining=zero-shotTraining=zero-shot2026.018.949.44UVISTraining=unsupervisedTraining=unsupervised2026.013.536.71VideoVae+Training=zero-shotTraining=zero-shot2026.012.829.91OmniTokenizerTraining=zero-shotTraining=zero-shot2026.012.833.27LARPTraining=zero-shotTraining=zero-shot2026.011.728.45