Image Classification on ImageNet 1k Depth (KNN & Linear Probing)
64.07Top-1 Accuracy (KNN)DINOv3
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| DINOv3Architecture=ViT-L/162026.01 | 64.07 | 83.48 | 69.15 | |
| DINOv3Architecture=ViT-L/162026.01 | 64.07 | 83.48 | 69.15 | |
| DeFMArchitecture=ViT-L/142026.01 | 63.46 | 84.79 | 71.72 | |
| DeFMArchitecture=ViT-L/142026.01 | 63.46 | 84.79 | 71.72 | |
| DINOv2Architecture=ViT-L/142026.01 | 61.93 | 81.76 | 68.7 | |
| DINOv2Architecture=ViT-L/142026.01 | 61.93 | 81.76 | 68.7 | |
| C-RADIOv3Architecture=ViT-L/16 (CPE)2026.01 | 55.9 | 77.39 | 64.12 | |
| C-RADIOv3Architecture=ViT-L/16 (CPE)2026.01 | 55.9 | 77.39 | 64.12 | |
| DeFMArchitecture=ViT-S/142026.01 | 55.27 | 78.06 | 61.54 | |
| DeFMArchitecture=ViT-S/142026.01 | 55.27 | 78.06 | 61.54 | |
| DINOv3Architecture=ViT-S/162026.01 | 49.5 | 70.93 | 50.86 | |
| DINOv3Architecture=ViT-S/162026.01 | 49.5 | 70.93 | 50.86 | |
| DINOv2Architecture=ViT-S/142026.01 | 47.66 | 69.11 | 55.06 | |
| DINOv2Architecture=ViT-S/142026.01 | 47.66 | 69.11 | 55.06 | |
| TheiaArchitecture=DeiT-S/162026.01 | 28.17 | 47.47 | 33.14 | |
| TheiaArchitecture=DeiT-S/162026.01 | 28.17 | 47.47 | 33.14 |