Classification on m-pv ger 4
98.3Overall Accuracy (OA)Dinov3 Web
Evaluation results
| Method | Links | |
|---|---|---|
| Dinov3 Webtraining modalities=Aerial, model=ViT-7B, linear probing=true2026.04 | 98.3 | |
| Dinov2training modalities=Aerial, model=ViT-B⋆, linear probing=true2026.04 | 97.5 | |
| DOFAtraining modalities=Aerial, model=ViT-L, linear probing=true2026.04 | 97.4 | |
| UniGeoCLIPtraining modalities=Aerial, GPS, Depth, Text, model=ViT-B, linear probing=true2026.04 | 97 | |
| UniGeoCLIPtraining modalities=Aerial, GPS, Depth, model=ViT-B, linear probing=true2026.04 | 96.9 | |
| Panopticontraining modalities=Aerial, GPS, model=ViT-B⋆, linear probing=true2026.04 | 96.4 | |
| SigLip2training modalities=Aerial, Text, model=ViT-B, linear probing=true2026.04 | 95.7 | |
| SatCliptraining modalities=Aerial, GPS, model=ViT-B, linear probing=true2026.04 | 93 | |
| AnySattraining modalities=Aerial, GPS, Depth, model=ViT-B, linear probing=true2026.04 | 92.8 | |
| SenPaMAEtraining modalities=Aerial, model=ViT-B, linear probing=true2026.04 | 87.1 |