Video Object Segmentation on ReVOS 1.0 (test)
48.8Jaccard IndexVISA
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| VISABackbone=LLaVA-13B, Instruction Tuning=true2024.07 | 48.8 | 52.8 | 50.8 | 15.1 | |
| VISABackbone=Chat-UniVi-13B, Instruction Tuning=true2024.07 | 48.8 | 52.9 | 50.9 | 14.5 | |
| VISABackbone=Chat-UniVi-13B, Instruction Tuning=false2024.07 | 45.3 | 49.7 | 47.5 | 8.3 | |
| VISABackbone=LLaVA-7B, Instruction Tuning=true2024.07 | 44.9 | 49.2 | 47.1 | 15.3 | |
| VISABackbone=Chat-UniVi-7B, Instruction Tuning=true2024.07 | 44.9 | 49 | 46.9 | 15.5 | |
| VISABackbone=Chat-UniVi-7B, Instruction Tuning=false2024.07 | 43.9 | 48.2 | 46.1 | 7.9 | |
| TrackGPTBackbone=LLaVA-13B, Instruction Tuning=true, Reproduced=true2024.07 | 43.2 | 46.8 | 45 | 12.8 | |
| TrackGPTBackbone=LLaVA-7B, Instruction Tuning=true, Reproduced=true2024.07 | 41.8 | 45.5 | 43.6 | 11.6 | |
| LISABackbone=LLaVA-13B, Instruction Tuning=false, Reproduced=true2024.07 | 39.8 | 43.5 | 41.6 | 8.6 | |
| LISABackbone=LLaVA-7B, Instruction Tuning=false2024.07 | 39.1 | 42.7 | 40.9 | 9.3 | |
| ReferFormerBackbone=Video-Swin-B, Instruction Tuning=false2024.07 | 26.2 | 29.9 | 28.1 | 8.8 | |
| MTTRBackbone=Video-Swin-T, Instruction Tuning=false2024.07 | 25.1 | 25.9 | 25.5 | 5.6 | |
| LMPMBackbone=Swin-T, Instruction Tuning=false2024.07 | 21.2 | 31.7 | 26.4 | 3.2 | |
| LLaMA-VID+LMPMBackbone=Swin-T, Instruction Tuning=false2024.07 | 20.9 | 31.4 | 26.1 | 3.4 | |
| ReferFormerBackbone=ResNet-50, Instruction Tuning=false2024.07 | 14.3 | 15.4 | 14.9 | 4.9 |