Referring Expression Segmentation on ReasonSeg (test)
64.2mIoUGETok-R1
Evaluation results
| Method | Links | |
|---|---|---|
| GETok-R1Training Mask Dec.=false, Learning Protocol=Reinforcement Learning2025.12 | 64.2 | |
| GETok-R1-gridTraining Mask Dec.=false, Learning Protocol=Reinforcement Learning2025.12 | 63.7 | |
| VisionReasonerTraining Mask Dec.=false, Learning Protocol=Reinforcement Learning2025.12 | 63.6 | |
| SAM-R1Training Mask Dec.=false, Learning Protocol=Reinforcement Learning2025.12 | 60.2 | |
| Seg-ZeroTraining Mask Dec.=false, Learning Protocol=Reinforcement Learning2025.12 | 57.5 | |
| GETok-SFTTraining Mask Dec.=false, Learning Protocol=Supervised Fine-Tuning2025.12 | 55.8 | |
| GETok-SFT-gridTraining Mask Dec.=false, Learning Protocol=Supervised Fine-Tuning2025.12 | 54.4 | |
| Qwen2.5-VL-7BTraining Mask Dec.=false, Learning Protocol=Supervised Fine-Tuning2025.12 | 51.5 | |
| LISATraining Mask Dec.=true, Learning Protocol=Supervised Fine-Tuning2025.12 | 36.8 |