Region-level Classification on NWPU-VHR-10
93.75Top-1 AccuracyGeoAlignCLIP
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| GeoAlignCLIPVersion=ViT-L/142026.03 | 93.75 | 99.97 | |
| GeoAlignCLIPVersion=ViT-B/162026.03 | 90.27 | 99.41 | |
| FG-CLIPVersion=ViT-L/14-3362026.03 | 88.37 | 99.86 | |
| FG-CLIPVersion=ViT-B/162026.03 | 86.16 | 98.27 | |
| SkyCLIPVersion=ViT-L/14-pct502026.03 | 83.64 | 98.76 | |
| SkyCLIPVersion=ViT-L/14-pct302026.03 | 81.82 | 98.41 | |
| VHMVersion=Vicuna-7B2026.03 | 75.69 | — | |
| SkysenseGPTVersion=Vicuna v1.5-7B2026.03 | 74.31 | — | |
| GeoRSCLIPVersion=ViT-L/142026.03 | 74.26 | 95.62 | |
| GeoRSCLIPVersion=ViT-B/322026.03 | 73.15 | 98.55 | |
| GeoRSCLIPVersion=ViT-L/14-3362026.03 | 72.77 | 91.55 | |
| LongCLIPVersion=ViT-L/142026.03 | 71.64 | 93.62 | |
| LRSCLIPVersion=ViT-L/142026.03 | 67.6 | 94.93 | |
| RemoteCLIPVersion=ViT-B/322026.03 | 66.01 | 91.41 | |
| LRSCLIPVersion=ViT-B/162026.03 | 64.8 | 95.96 | |
| LongCLIPVersion=ViT-B/162026.03 | 56.94 | 83.44 | |
| GeoChatVersion=Vicuna v1.5-7B2026.03 | 49.54 | — | |
| RemoteCLIPVersion=ViT-L/142026.03 | 15.25 | 42.79 | |
| CLIPVersion=ViT-B/162026.03 | 12.22 | 54.45 | |
| CLIPVersion=ViT-L/142026.03 | 11.87 | 56.9 | |
| SkyCLIPVersion=ViT-B/322026.03 | 8.32 | 67.18 |