Visual Grounding on ReasonSeg
61.6cIoU (Overall)ADPO
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| ADPOSample size=122026.01 | 61.6 | 53.9 | 56.2 | 67 | 61.3 | 62 | 73.6 | 60.9 | 73.2 | |
| ADPOSample size=82026.01 | 61.2 | 53.2 | 56 | 67 | 60.9 | 61.5 | 73.7 | 60.4 | 73.5 | |
| ADPOSample size=42026.01 | 61.1 | 52.2 | 55.1 | 67 | 61 | 61.5 | 73.3 | 60.5 | 72.9 | |
| GRPOSample size=82026.01 | 59.6 | 52 | 55.6 | 68 | 59.2 | 59.9 | 72 | 58.7 | 71.7 | |
| GRPOSample size=12026.01 | 59.5 | 51.8 | 55.5 | 67.9 | 59.1 | 59.7 | 71.3 | 58.6 | 71.1 | |
| GRPOSample size=42026.01 | 59.4 | 54.5 | 57 | 68 | 58.8 | 59.5 | 72.1 | 58.5 | 71.8 | |
| GRPOSample size=122026.01 | 59.4 | 55.6 | 58.1 | 69.9 | 58.8 | 59.5 | 72.2 | 58.6 | 72 | |
| ADPOSample size=12026.01 | 59.1 | 51.7 | 54.8 | 68 | 60.2 | 59.4 | 71.9 | 58.1 | 71.7 | |
| BaseSample size=122026.01 | 57.6 | 50.2 | 53.7 | 66 | 57.2 | 57.8 | 69.3 | 56.8 | 69.1 | |
| BaseSample size=42026.01 | 57.5 | 47.8 | 52 | 66 | 57.3 | 57.9 | 69.3 | 56.7 | 69.1 | |
| BaseSample size=82026.01 | 57.4 | 47.8 | 51.4 | 63.1 | 57.2 | 57.8 | 69.2 | 56.6 | 68.8 | |
| BaseSample size=12026.01 | 57.2 | 49.5 | 53 | 67 | 56.8 | 57.5 | 68.5 | 56.3 | 68.4 | |
| Seg-Zero-7BSample size=12026.01 | 52 | — | — | — | — | — | — | 57.5 | — | |
| LISA-7BSample size=12026.01 | 48.8 | 47.1 | 48.5 | — | 49.2 | 48.9 | — | 48.7 | — | |
| SegLLMSample size=12026.01 | 48.4 | — | — | — | — | 54.2 | — | — | — | |
| VLM-R1Sample size=12026.01 | — | — | — | — | — | — | — | — | 63.1 |