Classification on m-eurosat GEO-Bench
98.6Overall AccuracyTerraFM
Evaluation results
| Method | Links | |
|---|---|---|
| TerraFMBackbone=ViT-L2026.03 | 98.6 | |
| DeCURSetting=Fine-Tuning, Backbone=ViT-Small2026.05 | 97.9 | |
| HQ-JEPASetting=Fine-Tuning2026.05 | 97.81 | |
| GalileoBackbone=ViT-B2026.03 | 97.7 | |
| GeoSANEBackbone=ViT-L2026.03 | 97.7 | |
| DeCURSetting=Fine-Tuning, Backbone=Resnet502026.05 | 97.61 | |
| DOFABackbone=ViT-L2026.03 | 96.9 | |
| DINOSetting=Fine-Tuning, Backbone=Resnet502026.05 | 96.79 | |
| SatMAEBackbone=ViT-L2026.03 | 96.6 | |
| CROMABackbone=ViT-L2026.03 | 96.6 | |
| Prithvi-EO 2.0Backbone=ViT-L2026.03 | 96.5 | |
| Satlas-Swin-100MSetting=Fine-Tuning2026.05 | 96.28 | |
| DOFASetting=Fine-Tuning, Backbone=ViT-300M2026.05 | 96.18 | |
| AnySatBackbone=ViT-B2026.03 | 95.9 | |
| AnySatSetting=Fine-Tuning, Backbone=ViT-Base2026.05 | 95.9 | |
| Prithvi-EO-2.0-100MSetting=Fine-Tuning2026.05 | 94.72 | |
| ScaleMAE-ViT-300MSetting=Fine-Tuning2026.05 | 91.55 | |
| HQ-JEPASetting=Linear Probing2026.05 | 82.59 | |
| MAESetting=Linear Probing2026.05 | 78 | |
| SatMAESetting=Linear Probing2026.05 | 73.15 |