DINOv3 ViT-L/16
HuggingFace
GitHub
arXiv
Generic
Model Type
#2
Overall Rank
40.58%
Avg. R@1
29.73%
Avg.
mAP@20
mAP@20
1024
Embedding
Size
Size
224
Input
Size
Size
6
Datasets
Performance Across Datasets
| Dataset | Category | R@1 | R@5 | mAP |
|---|---|---|---|---|
| Stanford Online Products | E-commerce | 66.61% | 77.92% | 42.47% |
| Products-10K | E-commerce | 59.15% | 75.55% | 39.37% |
| DIY v1 | Hardware/DIY | 24.48% | 45.99% | 32.44% |
| Automotive v1 | Automotive | 26.29% | 47.28% | 28.45% |
| Clips-and-Connectors v1 | Industrial | 26.38% | 45.03% | 5.94% |
| Average | 40.58% | 58.35% | 29.73% | |
ILIAS
Showing the results for ILIAS separately, as it is excluded from the average calculations. Because not all models are evaluated on ILIAS.
| Dataset | Category | R@1 | R@5 | mAP@20 | mAP@1000 |
|---|---|---|---|---|---|
| ILIAS | Instance Retrieval | 47.97% | 56.74% | 21.82% | 22.40% |