VideoLoom: A Video Large Language Model for Joint Spatial-Temporal Understanding

Benchmark Model Rank Results
referring-expression-segmentation-on-refer-youtube-vos-2021-public-validationVideoLoom-8B (extra data)#6J&F: 71.3
referring-video-object-segmentation-on-revosVideoLoom-8B (extra data)#7J&F: 63.1
referring-video-object-segmentation-on-mevisVideoLoom-8B (extra data)#9J&F: 51.7
moment-retrieval-on-charades-staVideoLoom-8B#12R@1 IoU=0.5: 70.0R@1 IoU=0.7: 48.3