Image Classification on CIFAR-10 LT-10 (Robustness & OOD Evaluation)
90.65Clean AccuracyLA
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| LABackbone=WRN22-102025.01 | 90.65 | 69.73 | 91.14 | 4.59 | |
| HEMBackbone=WRN22-102025.01 | 89.95 | 71.37 | 94.81 | 4.72 | |
| HEM-Backbone=WRN22-102025.01 | 89.8 | 70.74 | 94.12 | 4.67 | |
| CEBackbone=WRN22-102025.01 | 89.74 | 69.02 | 86.52 | 4.64 | |
| MMBackbone=WRN22-102025.01 | 89.4 | 68.8 | 89.5 | 5.14 | |
| LABackbone=ResNet182025.01 | 89.36 | 70.6 | 86.51 | 5.29 | |
| LNBackbone=WRN22-102025.01 | 89.16 | 69.01 | 94.64 | 4.65 | |
| HEM-Backbone=ResNet182025.01 | 88.85 | 69.8 | 92.56 | 6.24 | |
| HEMBackbone=ResNet182025.01 | 88.74 | 69.81 | 90.31 | 6.38 | |
| CEBackbone=ResNet182025.01 | 88.52 | 69.59 | 85.27 | 5.25 | |
| MMBackbone=ResNet182025.01 | 88.51 | 68.73 | 88.3 | 5.54 | |
| LABackbone=ResNet322025.01 | 88.24 | 69.47 | 88.78 | 4.97 | |
| LNBackbone=ResNet182025.01 | 87.47 | 69.97 | 90.79 | 5.15 | |
| CEBackbone=ResNet322025.01 | 87.03 | 67.86 | 86.82 | 5.18 | |
| HEMBackbone=ResNet322025.01 | 86.9 | 69.36 | 91.59 | 5.91 | |
| LNBackbone=ResNet322025.01 | 86.81 | 67.71 | 91.96 | 5 | |
| HEM-Backbone=ResNet322025.01 | 86.73 | 68.45 | 91.15 | 6.22 | |
| MMBackbone=ResNet322025.01 | 86.61 | 67 | 85.68 | 5.13 | |
| DICEBackbone=WRN22-102025.01 | 85.92 | 64.81 | 87.54 | 5.75 | |
| DICEBackbone=ResNet182025.01 | 82.25 | 63.05 | 82.85 | 6.65 | |
| DICEBackbone=ResNet322025.01 | 80.02 | 60.33 | 79.58 | 7.53 |