Adversarial Robustness on STL10 (test)
82.07Baseline AccZePAD
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| ZePADPre-trained Model=SimCLR2026.02 | 82.07 | 77.56 | 76.51 | 72.93 | 72.1 | 73.26 | |
| ZePADPre-trained Model=MoCo v2+2026.02 | 81.6 | 76.05 | 74.83 | 72.15 | 67.96 | 74.3 | |
| Gen-AFPre-trained Model=MoCo v2+2026.02 | 69.09 | 64.98 | 68.15 | 59.41 | 60.47 | 66 | |
| Gen-AFPre-trained Model=SimCLR2026.02 | 68.69 | 64.43 | 68.65 | 65.92 | 65.48 | 62.28 | |
| TRADESPre-trained Model=SimCLR2026.02 | 50.79 | 50.4 | 50.51 | 50.95 | 50.59 | 50.66 | |
| TRADESPre-trained Model=MoCo v2+2026.02 | 50.5 | 52.4 | 50.6 | 52.37 | 52.6 | 52.28 | |
| MARTPre-trained Model=SimCLR2026.02 | 42.79 | 42.5 | 42.65 | 42.6 | 42.24 | 42.36 | |
| MARTPre-trained Model=MoCo v2+2026.02 | 42.79 | 41.46 | 41.52 | 41.28 | 41.76 | 41.36 | |
| PGD-ATPre-trained Model=SimCLR2026.02 | 22.19 | 21.23 | 23 | 21.79 | 20.97 | 19.04 | |
| PGD-ATPre-trained Model=MoCo v2+2026.02 | 13.72 | 12.15 | 13.48 | 12.3 | 13.02 | 12.7 |