Adversarial Defense On Imagenet Non Targeted
Metrics
Accuracy
Results
Performance results of various models on this benchmark
Comparison Table
| Model Name | Accuracy |
|---|---|
| adversarial-training-for-free | 36.0% |
| adversarial-training-for-free | 34.3% |
| adversarial-robustness-through-local | 47.0% |
| smooth-adversarial-training | 58.6% |
| adversarial-training-for-free | 31.8% |