Image Classification on CIFAR-10 (Certified Robustness vs Perturbation Level)
92.7Clean AccuracyLipNeXt L32W2896
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| LipNeXt L32W2896#Param.=512M2026.01 | 92.7 | 81.7 | 68.6 | 55.8 | |
| LipNeXt L32W2048#Param.=256M2026.01 | 88.2 | 79.2 | 68 | 54.9 | |
| BRO#Param.=68M2026.01 | 87.2 | 78.3 | 67.4 | 54.5 | |
| Cayley#Param.=68M2026.01 | 86.7 | 77.7 | 66.9 | 54.3 | |
| Lip-reg#Param.=83M2026.01 | 86.7 | 78.1 | 67 | 54.2 | |
| SOC#Param.=83M2026.01 | 86.6 | 78.2 | 67 | 54.1 | |
| LOT#Param.=59M2026.01 | 85.7 | 76.4 | 65.1 | 52.2 | |
| SLL#Param.=83M2026.01 | 85.6 | 76.8 | 66 | 53.3 | |
| Cholesky#Param.=68M2026.01 | 85.4 | 76.6 | 65.7 | 53.3 |