Image Classification on ImageNet (Top-1 Clean and Robust Accuracy)
68.42Top-1 Clean AccuracyCLBP
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| CLBPShots=16, Training Epochs=10, Training Attack (PGD)=2-step (1/255), Evaluation Attack (PGD-100)=100-step (1/255)2026.05 | 68.42 | 59.88 | |
| AdvVLPShots=16, Training Epochs=10, Training Attack (PGD)=2-step (1/255), Evaluation Attack (PGD-100)=100-step (1/255)2026.05 | 53.23 | 22.1 | |
| AdvMaPLeShots=16, Training Epochs=10, Training Attack (PGD)=2-step (1/255), Evaluation Attack (PGD-100)=100-step (1/255)2026.05 | 52.93 | 21.9 | |
| FAPShots=16, Training Epochs=10, Training Attack (PGD)=2-step (1/255), Evaluation Attack (PGD-100)=100-step (1/255)2026.05 | 52.53 | 22.9 | |
| AdvVPShots=16, Training Epochs=10, Training Attack (PGD)=2-step (1/255), Evaluation Attack (PGD-100)=100-step (1/255)2026.05 | 46.27 | 12.77 |