Image Classification on CIFAR10 32/255 noise (test)
70.84Natural AccuracyMultiGrad
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MultiGradBackbone=PreActResNet-182024.04 | 70.84 | 0 | |
| ZeroGradBackbone=PreActResNet-182024.04 | 65.48 | 0 | |
| FreeATBackbone=PreActResNet-182024.04 | 61.11 | 0 | |
| PGD-2Backbone=PreActResNet-182024.04 | 48.45 | 11.24 | |
| RS-FGSMBackbone=PreActResNet-182024.04 | 36.43 | 0 | |
| PGD-10(20)Backbone=PreActResNet-182024.04 | 34.7 | 16.1 | |
| RS-AAERBackbone=PreActResNet-182024.04 | 31.58 | 10.62 | |
| N-FGSMBackbone=PreActResNet-182024.04 | 29.79 | 8.3 | |
| N-AAERBackbone=PreActResNet-182024.04 | 27.08 | 12.97 | |
| Grad AlignBackbone=PreActResNet-182024.04 | 25.23 | 11.53 |