Image Classification on fMoW-RGB
69.2Top-1 AccCross-Scale MAE
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Cross-Scale MAEBackbone=ViT-Base, Scale=100%, Evaluation Protocol=Linear classification fine-tuning2024.01 | 69.2 | 95.4 | |
| GASSLBackbone=ResNet50, Scale=100%, Evaluation Protocol=Linear classification fine-tuning2024.01 | 68.3 | 71.5 | |
| SatMAEBackbone=ViT-Large, Scale=100%, Evaluation Protocol=Linear classification fine-tuning2024.01 | 67.8 | 92.3 | |
| Scale-MAEBackbone=ViT-Base, Scale=100%, Evaluation Protocol=Linear classification fine-tuning2024.01 | 67.3 | 93 | |
| SatMAEBackbone=ViT-Base, Scale=100%, Evaluation Protocol=Linear classification fine-tuning2024.01 | 65.1 | 89.2 | |
| Cross-Scale MAEBackbone=ViT-Base, Scale=50%, Evaluation Protocol=Linear classification fine-tuning2024.01 | 63.2 | 91.4 | |
| Scale-MAEBackbone=ViT-Base, Scale=50%, Evaluation Protocol=Linear classification fine-tuning2024.01 | 62.3 | 85 | |
| SeCoBackbone=ResNet50, Scale=100%, Evaluation Protocol=Linear classification fine-tuning2024.01 | 61.4 | 79.6 | |
| SatMAEBackbone=ViT-Large, Scale=50%, Evaluation Protocol=Linear classification fine-tuning2024.01 | 59.1 | 79.3 | |
| SatMAEBackbone=ViT-Base, Scale=50%, Evaluation Protocol=Linear classification fine-tuning2024.01 | 55.1 | 77.6 | |
| SeCoBackbone=ResNet50, Scale=50%, Evaluation Protocol=Linear classification fine-tuning2024.01 | 50.8 | 60.2 | |
| GASSLBackbone=ResNet50, Scale=50%, Evaluation Protocol=Linear classification fine-tuning2024.01 | 49.4 | 54 |