Referring Expression Grounding on RefCOCOg
88AccuracyQwen-2.5-VL-7B + EvoLMM
Evaluation results
| Method | Links | |
|---|---|---|
| Qwen-2.5-VL-7B + EvoLMMsetting=zero-shot2025.11 | 88 | |
| Qwen-2.5-VL-7Bsetting=zero-shot2025.11 | 87.2 | |
| WUDI Mergingmode=Full fine-tuning, backbone=InternVL2.5-1B2026.06 | 74.48 | |
| OptMergemode=Full fine-tuning, backbone=InternVL2.5-1B2026.06 | 74.18 | |
| SWUDImode=Full fine-tuning, backbone=InternVL2.5-1B2026.06 | 74.12 | |
| SWUDI-Amode=Full fine-tuning, backbone=InternVL2.5-1B2026.06 | 74.08 | |
| TIES Mergingmode=Full fine-tuning, backbone=InternVL2.5-1B2026.06 | 73.65 | |
| TIES w/ DAREmode=Full fine-tuning, backbone=InternVL2.5-1B2026.06 | 73.55 | |
| TA w/ DAREmode=Full fine-tuning, backbone=InternVL2.5-1B2026.06 | 73.02 | |
| Task Arithmeticmode=Full fine-tuning, backbone=InternVL2.5-1B2026.06 | 72.75 | |
| Iso-Cmode=Full fine-tuning, backbone=InternVL2.5-1B2026.06 | 68.5 | |
| Mixture Trainingmode=Full fine-tuning, backbone=InternVL2.5-1B2026.06 | 67.46 | |
| InternVL2.5-Instructmode=Full fine-tuning, backbone=InternVL2.5-1B2026.06 | 67.4 | |
| TSV Mergingmode=Full fine-tuning, backbone=InternVL2.5-1B2026.06 | 59.17 | |
| Weight Averagemode=Full fine-tuning, backbone=InternVL2.5-1B2026.06 | 52.39 |