Image Classification on CIFAR100 32/255 noise (test)
49.38Natural AccuracyZeroGrad
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ZeroGradBackbone=PreActResNet-182024.04 | 49.38 | 0 | |
| PGD-2Backbone=PreActResNet-182024.04 | 31.89 | 4.51 | |
| MultiGradBackbone=PreActResNet-182024.04 | 28.33 | 0 | |
| FreeATBackbone=PreActResNet-182024.04 | 26.2 | 0 | |
| PGD-10(20)Backbone=PreActResNet-182024.04 | 21.68 | 7.39 | |
| RS-AAERBackbone=PreActResNet-182024.04 | 18.5 | 4.9 | |
| N-FGSMBackbone=PreActResNet-182024.04 | 18.18 | 0 | |
| N-AAERBackbone=PreActResNet-182024.04 | 16.95 | 5.45 | |
| Grad AlignBackbone=PreActResNet-182024.04 | 15.05 | 2.85 | |
| RS-FGSMBackbone=PreActResNet-182024.04 | 11.4 | 0 |