Out-of-Distribution Detection on ImageNet-1k vs SSB-Hard
77.28AUROCDINOv2+MLS
Evaluation results
| Method | Links | ||
|---|---|---|---|
| DINOv2+MLSBackbone=DINOv2, Training=Linear Probe, Postprocessor=MLS2024.10 | 77.28 | 72.9 | |
| ViT-B+CE+RMDSBackbone=ViT-B, Training=Cross Entropy (CE), Postprocessor=RMDS2024.10 | 72.87 | 84.52 | |
| GENBackbone=Swin-T, ID Accuracy=81.60%2026.04 | 72.78 | 79.35 | |
| TempScaleBackbone=Swin-T, ID Accuracy=81.60%2026.04 | 71.83 | 82.12 | |
| RMDSBackbone=Swin-T, ID Accuracy=81.60%2026.04 | 71.81 | 82.66 | |
| MSPBackbone=Swin-T, ID Accuracy=81.60%2026.04 | 71.75 | 81.02 | |
| OpenMaxBackbone=Swin-T, ID Accuracy=81.60%2026.04 | 71.52 | 85.54 | |
| SHEBackbone=Swin-T, ID Accuracy=81.60%2026.04 | 70.75 | 85.03 | |
| Energy Profile Divergence (EPD)Backbone=Swin-T, ID Accuracy=81.60%2026.04 | 70.5 | 81.62 | |
| MLSBackbone=Swin-T, ID Accuracy=81.60%2026.04 | 70.47 | 86.6 | |
| ReActBackbone=Swin-T, ID Accuracy=81.60%2026.04 | 69.36 | 85.01 | |
| ViMBackbone=Swin-T, ID Accuracy=81.60%2026.04 | 68.94 | 88.55 | |
| MDSBackbone=Swin-T, ID Accuracy=81.60%2026.04 | 68.69 | 83.72 | |
| EBOBackbone=Swin-T, ID Accuracy=81.60%2026.04 | 68.28 | 87.51 | |
| KNNBackbone=Swin-T, ID Accuracy=81.60%2026.04 | 64.21 | 85.08 | |
| GradNormBackbone=Swin-T, ID Accuracy=81.60%2026.04 | 50.43 | 93.37 | |
| DICEBackbone=Swin-T, ID Accuracy=81.60%2026.04 | 49.97 | 95.76 | |
| ASHBackbone=Swin-T, ID Accuracy=81.60%2026.04 | 46.28 | 95.22 |