Video Reasoning Segmentation on VideoReasonSeg Long (val)
42Seg J&FViLLa
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ViLLaBackbone=InternVideo2-6B2024.07 | 42 | 41 | |
| ViLLaBackbone=InternVideo2-1B2024.07 | 38.9 | 37.6 | |
| VISABackbone=InternVideo2-6B2024.07 | 36.2 | 35.8 | |
| VISABackbone=Chat-UniVi-13B2024.07 | 34 | 34.3 | |
| VideoLISA*Backbone=InternVideo2-6B2024.07 | 33.8 | 34.9 | |
| VISABackbone=LLaVA-13B2024.07 | 33.8 | 33.2 | |
| VideoLISA*Backbone=InternVideo2-1B2024.07 | 32.4 | 32.9 | |
| VISABackbone=LLaVA-7B2024.07 | 30.4 | 31.9 | |
| LISABackbone=LLaVA-13B2024.07 | 28.1 | 30.7 | |
| LISABackbone=LLaVA-7B2024.07 | 27.3 | 29.5 | |
| OnlineRefer2024.07 | 18.7 | — | |
| LMPM2024.07 | 18.5 | — | |
| ReferFormer2024.07 | 13.4 | — | |
| MTTR2024.07 | 12.7 | — |