Image Classification on CIFAR-10 (test) (Accuracy and Overhead Profile)
77.88Clean AccuracyVanilla
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| VanillaBackbone=ResNet-18, TASER regularizer=false2026.05 | 77.88 | 3 | — | |
| VanillaBackbone=ResNet-18, TASER regularizer=true2026.05 | 70.92 | 19.25 | 2.45 | |
| TRADES + AWPBackbone=ResNet-18, TASER regularizer=true2026.05 | 68.87 | 36.2 | 1.21 | |
| TRADES + AWPBackbone=ResNet-18, TASER regularizer=false2026.05 | 67.9 | 34.05 | — | |
| TRADESBackbone=ResNet-18, TASER regularizer=false2026.05 | 67.07 | 27.35 | — | |
| MART + AWPBackbone=ResNet-18, TASER regularizer=false2026.05 | 66.96 | 32.3 | — | |
| MART + AWPBackbone=ResNet-18, TASER regularizer=true2026.05 | 66.14 | 37.4 | 1.17 | |
| PGDBackbone=ResNet-18, TASER regularizer=false2026.05 | 66.03 | 24.25 | — | |
| PGDBackbone=ResNet-18, TASER regularizer=true2026.05 | 65.64 | 33.35 | 1.25 | |
| TRADESBackbone=ResNet-18, TASER regularizer=true2026.05 | 65.61 | 33.95 | 1.27 | |
| MARTBackbone=ResNet-18, TASER regularizer=false2026.05 | 64.07 | 26.35 | — | |
| MARTBackbone=ResNet-18, TASER regularizer=true2026.05 | 63.31 | 35 | 1.19 |