Image Classification on CIFAR10 16/255 noise (test)
82.98Natural AccuracyMultiGrad
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MultiGradBackbone=PreActResNet-182024.04 | 82.98 | 0 | |
| ZeroGradBackbone=PreActResNet-182024.04 | 79.65 | 6.37 | |
| PGD-2Backbone=PreActResNet-182024.04 | 72.31 | 24.32 | |
| PGD-10(20)Backbone=PreActResNet-182024.04 | 67.2 | 29.34 | |
| RS-FGSMBackbone=PreActResNet-182024.04 | 66.54 | 0 | |
| RS-AAERBackbone=PreActResNet-182024.04 | 64.56 | 23.87 | |
| N-FGSMBackbone=PreActResNet-182024.04 | 62.96 | 27.14 | |
| N-AAERBackbone=PreActResNet-182024.04 | 61.84 | 28.2 | |
| Grad AlignBackbone=PreActResNet-182024.04 | 60.37 | 27.9 | |
| FreeATBackbone=PreActResNet-182024.04 | 45.84 | 0 |