Image Classification on STL-10 (test) (Adversarial Robustness)
52.14Accuracy (Benign)Lorica
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| LoricaBackbone=ViT-L/162025.06 | 52.14 | 47.22 | 45.1 | 47.22 | 46.39 | |
| SylvaBackbone=ViT-L/162025.06 | 50.51 | 45.01 | 43.52 | 46.47 | 45.42 | |
| FP-Gen-AFBackbone=ViT-L/162025.06 | 47.2 | 41.03 | 40.49 | 42.12 | 43.02 | |
| Per-LoRABackbone=ViT-L/162025.06 | 46.95 | 41.78 | 41.2 | 42.05 | 42.11 | |
| DBFATBackbone=ViT-L/162025.06 | 46.79 | 42.17 | 39.91 | 41.14 | 42.78 | |
| FA-Gen-AFBackbone=ViT-L/162025.06 | 46 | 40.48 | 39.12 | 41.98 | 41.85 | |
| FP-TRADESBackbone=ViT-L/162025.06 | 45.02 | 40.72 | 39.12 | 40.93 | 40.99 | |
| FA-TRADESBackbone=ViT-L/162025.06 | 43.7 | 39.15 | 38.11 | 41.8 | 40.93 | |
| Per-AdvBackbone=ViT-L/162025.06 | 37.95 | 41.23 | 39.48 | 41.52 | 41.12 | |
| FP-PGD-ATBackbone=ViT-L/162025.06 | 37.78 | 41.77 | 39.14 | 42.51 | 41.69 | |
| FA-PGD-ATBackbone=ViT-L/162025.06 | 35.89 | 41.05 | 39.88 | 41.95 | 41.05 |