List your company
ResearchFormerly WizWand

Image Classification on ImageNet-1k (Adversarial Robustness)

56.29Clean AccuracyRS-FGSM
47.210849.567951.92554.2821May 5, 2025
Updated 5mo ago

Evaluation results

MethodLinks
RS-FGSMBackbone=ResNet-50, Epsilon=2/255, Attack Protocol=PGD-50-10, Training Epochs=15, Pre-training=ImageNet-1k2025.0556.2936.86
FGSMBackbone=ResNet-50, Epsilon=2/255, Attack Protocol=PGD-50-10, Training Epochs=15, Pre-training=ImageNet-1k2025.0554.7238.21
N-FGSMBackbone=ResNet-50, Epsilon=2/255, Attack Protocol=PGD-50-10, Training Epochs=15, Pre-training=ImageNet-1k2025.0554.3938.07
lp-FGSMBackbone=ResNet-50, Epsilon=2/255, Attack Protocol=PGD-50-10, Training Epochs=15, Pre-training=ImageNet-1k2025.0553.1837.94
RS-FGSMBackbone=ResNet-50, Epsilon=4/255, Attack Protocol=PGD-50-10, Training Epochs=15, Pre-training=ImageNet-1k2025.0550.8125.12
lp-FGSMBackbone=ResNet-50, Epsilon=6/255, Attack Protocol=PGD-50-10, Training Epochs=15, Pre-training=ImageNet-1k2025.0548.6119.79
FGSMBackbone=ResNet-50, Epsilon=6/255, Attack Protocol=PGD-50-10, Training Epochs=15, Pre-training=ImageNet-1k2025.0548.550.08
FGSMBackbone=ResNet-50, Epsilon=4/255, Attack Protocol=PGD-50-10, Training Epochs=15, Pre-training=ImageNet-1k2025.0548.525.86
lp-FGSMBackbone=ResNet-50, Epsilon=4/255, Attack Protocol=PGD-50-10, Training Epochs=15, Pre-training=ImageNet-1k2025.0548.4228.35
N-FGSMBackbone=ResNet-50, Epsilon=6/255, Attack Protocol=PGD-50-10, Training Epochs=15, Pre-training=ImageNet-1k2025.0547.717.12
RS-FGSMBackbone=ResNet-50, Epsilon=6/255, Attack Protocol=PGD-50-10, Training Epochs=15, Pre-training=ImageNet-1k2025.0547.6716.49
N-FGSMBackbone=ResNet-50, Epsilon=4/255, Attack Protocol=PGD-50-10, Training Epochs=15, Pre-training=ImageNet-1k2025.0547.5626.28