| adversarial-robustness-on-imagenet-c | DINOv3 (7B) | #1 | mean Corruption Error (mCE): 19.6 |
| frozen-feature-label-propagation-on-davis-2017-val | DINOv3 + label propagation | #1 | J&F: 83.3 |
| image-retrieval-on-amstertime | DINOv3 (7B) | #1 | mAP: 56.5 |
| linear-probe-image-classification-on-food-101 | DINOv3 (7B) | #1 | Accuracy: 96.9 |
| linear-probe-semantic-segmentation-on-ade20k | DINOv3 + linear probe | #1 | mIoU: 55.9 |
| object-detection-on-coco-val2017 | DINOv3 7B/16 (Plain-DETR, frozen backbone, TTA) | #1 | box AP: 66.1Params (M): 7100 |
| object-detection-on-coco-val2017 | DINOv3 7B/16 (Plain-DETR, frozen backbone, no TTA) | #4 | box AP: 65.6Params (M): 7100 |
| image-classification-on-places205 | DINOv3 (7B) | #2 | Top 1 Accuracy: 70.0 |
| linear-probe-image-classification-on-sun397 | DINOv3 (7B) | #3 | Accuracy: 81.1 |
| monocular-depth-estimation-on-kitti-relative | DINOv3 + Depth Anything V2 | #3 | AbsRel: 0.073delta1: 0.967 |
| monocular-depth-estimation-on-nyuv2-relative | DINOv3 + Depth Anything V2 | #3 | delta1: 0.98AbsRel: 0.043 |
| monocular-depth-estimation-on-scannet-relative-depth-aligned | DINOv3 + Depth Anything V2 | #3 | AbsRel: 0.044delta1: 0.981 |
| linear-probe-image-classification-on-caltech-101 | DINOv3 (7B) | #4 | Accuracy: 94.8 |
| monocular-depth-estimation-on-eth3d-relative-depth-aligned | DINOv3 + Depth Anything V2 | #4 | AbsRel: 0.054delta1: 0.975 |
| relative-pose-estimation-on-scannet-1500 | DINOv3 + VGGT | #8 | AUC@10: 56.1 |
| image-classification-on-objectnet | DINOv3 (7B) | #9 | Top-1 Accuracy: 79.0 |
| image-classification-on-imagenet-v2 | DINOv3 (7B) | #11 | Top 1 Accuracy: 81.4 |
| image-classification-on-imagenet-real | DINOv3 (7B) | #22 | Accuracy: 90.4% |
| 3d-reconstruction-on-dtu | DINOv3 + VGGT | #25 | Overall: 0.368 |
| image-classification-on-imagenet | DINOv3 (7B) | #53 | Top 1 Accuracy: 88.4% |