Retrieval on MT-Popular Cities (All Queries)
58.19Recall@5SMGR
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| SMGRtext_encoder=Qwen3-0.6B, image_encoder=cn-CLIP-ViT-h-142026.02 | 58.19 | 41.65 | 72.92 | 44.02 | 84.1 | 48.16 | — | — | — | |
| MTIGERtext_encoder=Qwen3-0.6B, image_encoder=cn-CLIP-ViT-h-142026.02 | 56.01 | 39.58 | 71.18 | 42.08 | 82.42 | 47.17 | — | — | — | |
| TIGERtext_encoder=Qwen3-0.6B, image_encoder=cn-CLIP-ViT-h-142026.02 | 53.82 | 37.3 | 69.49 | 40.31 | 80.74 | 45.45 | — | — | — | |
| Joint-Que2searchtext_encoder=Qwen3-0.6B, image_encoder=cn-CLIP-ViT-h-142026.02 | 53.78 | 37.22 | 69.46 | 40.26 | 80.73 | 45.41 | — | — | — | |
| Qwen3-DualTowertext_encoder=Qwen3-0.6B, image_encoder=cn-CLIP-ViT-h-142026.02 | 53.48 | 36.98 | 68.99 | 40.17 | 80.46 | 45.34 | — | — | — | |
| SMGRtext_encoder=Qwen3-0.6B, image_encoder=cn-CLIP-ViT-h-142026.02 | 51.12 | — | 66.74 | — | 79.5 | — | 35.09 | 38.42 | 42.39 | |
| MTIGERtext_encoder=Qwen3-0.6B, image_encoder=cn-CLIP-ViT-h-142026.02 | 47.91 | — | 63.68 | — | 77.58 | — | 32.98 | 36.45 | 41.2 | |
| TIGERtext_encoder=Qwen3-0.6B, image_encoder=cn-CLIP-ViT-h-142026.02 | 45.38 | — | 61.79 | — | 75.99 | — | 31.02 | 34.88 | 40.14 | |
| Joint-Que2searchtext_encoder=Qwen3-0.6B, image_encoder=cn-CLIP-ViT-h-142026.02 | 45.05 | — | 61.5 | — | 75.87 | — | 30.66 | 34.52 | 39.91 | |
| Qwen3-DualTowertext_encoder=Qwen3-0.6B, image_encoder=cn-CLIP-ViT-h-142026.02 | 44.68 | — | 61.15 | — | 75.74 | — | 30.27 | 34.38 | 39.81 |