Image Classification on DTD PGD-l∞ attack, ε=4/255 (test)
49.43Robust AccuracyCore-tuning
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Core-tuningBackbone=ResNet-50, Pre-training=MoCo-v2, Attack Type=PGD-10, Perturbation Norm=l∞, Epsilon=4/2552021.02 | 49.43 | 70.65 | |
| AT-CE-ConBackbone=ResNet-50, Pre-training=MoCo-v2, Attack Type=PGD-10, Perturbation Norm=l∞, Epsilon=4/2552021.02 | 48.95 | 69.07 | |
| AT-CEBackbone=ResNet-50, Pre-training=MoCo-v2, Attack Type=PGD-10, Perturbation Norm=l∞, Epsilon=4/2552021.02 | 47.27 | 67.13 | |
| CEBackbone=ResNet-50, Pre-training=MoCo-v2, Attack Type=PGD-10, Perturbation Norm=l∞, Epsilon=4/2552021.02 | 18.37 | 71.68 |