Multi-modal Task Evaluation on MM-MergeBench Unseen Tasks
79.3AVQA ScoreRobustMerge
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| RobustMergeSetting=Unseen/Zero-shot generalization2026.03 | 79.3 | 45.79 | 9.23 | 17.62 | 37.99 | |
| DC-MergeSetting=Unseen/Zero-shot generalization2026.03 | 79.3 | 50.24 | 10.88 | 18.94 | 39.84 | |
| Iso-CTSSetting=Unseen/Zero-shot generalization2026.03 | 78.48 | 44.81 | 11.21 | 13.99 | 37.12 | |
| TSV-MSetting=Unseen/Zero-shot generalization2026.03 | 77.29 | 42.72 | 11.12 | 13.8 | 36.23 | |
| Multi TaskSetting=Unseen/Zero-shot generalization2026.03 | 76.33 | 41.39 | 8.34 | 18.2 | 36.06 | |
| Task ArithmeticSetting=Unseen/Zero-shot generalization2026.03 | 74.78 | 37.37 | 7.52 | 13.57 | 33.31 | |
| PCB-MergingSetting=Unseen/Zero-shot generalization2026.03 | 74.57 | 36.28 | 7.84 | 15.44 | 33.53 | |
| DARE-MergingSetting=Unseen/Zero-shot generalization2026.03 | 73.75 | 37.67 | 7.56 | 13.62 | 33.15 | |
| TIES-MergingSetting=Unseen/Zero-shot generalization2026.03 | 73.43 | 38.44 | 7.47 | 13.23 | 33.14 | |
| WUDI-MergingSetting=Unseen/Zero-shot generalization2026.03 | 68.56 | 41.57 | 11.16 | 11.8 | 33.27 | |
| ZeroshotSetting=Unseen/Zero-shot generalization2026.03 | 51.62 | 28.27 | 5.98 | 15.01 | 25.22 |