Selective Classification on Average Covariate Shift
39.9Metric AMDS
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MDSBackbone=ResNet-50, Training Dataset=ImageNet-1K, Pre-training Source=Official PyTorch Repository2025.05 | 39.9 | 97.7 | |
| KNNBackbone=ResNet-50, Training Dataset=ImageNet-1K, Pre-training Source=Official PyTorch Repository2025.05 | 36.4 | 94 | |
| SIRCBackbone=ResNet-50, Training Dataset=ImageNet-1K, Pre-training Source=Official PyTorch Repository2025.05 | 31.3 | 71.8 | |
| Δ-KNNBackbone=ResNet-50, Training Dataset=ImageNet-1K, Pre-training Source=Official PyTorch Repository2025.05 | 25.3 | 52.1 | |
| EnergyBackbone=ResNet-50, Training Dataset=ImageNet-1K, Pre-training Source=Official PyTorch Repository2025.05 | 22.5 | 44.8 | |
| Δ-MDSBackbone=ResNet-50, Training Dataset=ImageNet-1K, Pre-training Source=Official PyTorch Repository2025.05 | 21.1 | 38 | |
| MaxLogitBackbone=ResNet-50, Training Dataset=ImageNet-1K, Pre-training Source=Official PyTorch Repository2025.05 | 19 | 32 | |
| MSPBackbone=ResNet-50, Training Dataset=ImageNet-1K, Pre-training Source=Official PyTorch Repository2025.05 | 16.5 | 22.6 | |
| RLogBackbone=ResNet-50, Training Dataset=ImageNet-1K, Pre-training Source=Official PyTorch Repository2025.05 | 16.2 | 20.7 | |
| Δ-KNN-RLogBackbone=ResNet-50, Training Dataset=ImageNet-1K, Pre-training Source=Official PyTorch Repository2025.05 | 16 | 20.2 | |
| Δ-MDS-RLogBackbone=ResNet-50, Training Dataset=ImageNet-1K, Pre-training Source=Official PyTorch Repository2025.05 | 15.8 | 19.5 |