Classification on Resisc45
94.85Top-1 AccuracySkySense
Evaluation results
| Method | Links | |
|---|---|---|
| SkySenseBackbone=Swin V2-H, Training Ratio=10%2026.04 | 94.85 | |
| CoDe-MAEBackbone=HiViT-B, Training Ratio=10%2026.04 | 94.79 | |
| SelectiveMAEBackbone=ViT-L, Training Ratio=10%2026.04 | 94.57 | |
| SkySenseBackbone=Swin V2-L, Training Ratio=10%2026.04 | 94.34 | |
| RingMoBackbone=Swin-B, Training Ratio=10%2026.04 | 94.25 | |
| RVSABackbone=ViT-B+RVSA, Training Ratio=10%2026.04 | 93.93 | |
| SelectiveMAEBackbone=ViT-B, Training Ratio=10%2026.04 | 93.35 | |
| MaRSBackbone=Swin V2-B, Training Ratio=10%2026.04 | 92.87 | |
| GFMBackbone=Swin-B, Training Ratio=10%2026.04 | 92.73 | |
| ScaleMAEBackbone=ViT-L, Training Ratio=10%2026.04 | 92.63 | |
| SatLasBackbone=Swin-B, Training Ratio=10%2026.04 | 92.16 | |
| SatMAEBackbone=ViT-L, Training Ratio=10%2026.04 | 91.72 | |
| ITOPre-training Dataset=DataComp-1B, Backbone=ViT-B/16, Training Epochs=10 epochs, Zero-shot protocol=EVA-CLIP2026.03 | 69.9 | |
| ITO sub2Pre-training Dataset=DataComp-1B, Backbone=ViT-B/16, Training Epochs=10 epochs, Zero-shot protocol=EVA-CLIP2026.03 | 66.7 | |
| CLIPPre-training Dataset=DataComp-1B, Backbone=ViT-B/16, Training Epochs=10 epochs, Zero-shot protocol=EVA-CLIP2026.03 | 66.1 | |
| ITO sub2Backbone=ViT-B/16, Pre-training Dataset=CC3M-recap, Evaluation Protocol=Zero-shot2026.03 | 37.3 | |
| ITO sub3Backbone=ViT-B/16, Pre-training Dataset=CC3M-recap, Evaluation Protocol=Zero-shot2026.03 | 35 | |
| FLAIRBackbone=ViT-B/16, Pre-training Dataset=CC3M-recap, Evaluation Protocol=Zero-shot2026.03 | 34.6 |