Classification on CIFAR-10-C 1.0 (test)
88.02Accuracy (Clean)MGR
Evaluation results
| Method | Links | |||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| MGRBackbone=ResNet-182023.07 | 88.02 | 61.69 | 62.53 | 41.62 | 52.21 | 51.91 | 47.94 | 51.09 | 65.78 | 59.5 | 34.49 | 73.23 | 14.83 | 60.65 | 65.32 | 71.1 | 56.37 | |
| PCRBackbone=ResNet-182023.07 | 87.06 | 59.35 | 60.65 | 37.71 | 47.51 | 50.89 | 41.71 | 47.11 | 63.94 | 57.33 | 31.13 | 73.43 | 13.66 | 58 | 63.77 | 68.55 | 53.86 | |
| Base ModelBackbone=ResNet-182023.07 | 86.49 | 53.32 | 53.12 | 39.01 | 34.06 | 37.15 | 31.08 | 36.14 | 46.68 | 36.38 | 19.14 | 52.45 | 10.51 | 41.14 | 46.59 | 53.02 | 42.27 | |
| GDABackbone=ResNet-182023.07 | 84.11 | 52.7 | 52.98 | 39.94 | 29.39 | 38.66 | 28.97 | 28.82 | 45.04 | 40.08 | 24 | 50.29 | 13.07 | 40.91 | 45.11 | 52.85 | 41.68 |