Image Classification on CIFAR100 16/255 noise (test)
50.76Natural AccuracyZeroGrad
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ZeroGradBackbone=PreActResNet-182024.04 | 50.76 | 5.23 | |
| PGD-2Backbone=PreActResNet-182024.04 | 47.5 | 12.69 | |
| MultiGradBackbone=PreActResNet-182024.04 | 46.05 | 0 | |
| PGD-10(20)Backbone=PreActResNet-182024.04 | 42.39 | 15.01 | |
| N-FGSMBackbone=PreActResNet-182024.04 | 37.93 | 14.05 | |
| Grad AlignBackbone=PreActResNet-182024.04 | 36.8 | 13.77 | |
| N-AAERBackbone=PreActResNet-182024.04 | 36.8 | 14.31 | |
| RS-AAERBackbone=PreActResNet-182024.04 | 33.1 | 11.8 | |
| RS-FGSMBackbone=PreActResNet-182024.04 | 11.03 | 0 | |
| FreeATBackbone=PreActResNet-182024.04 | 3.32 | 0 |