Image Classification on ImageNet large-scale (test)
32.15Verified Error (%)ABSCERT
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ABSCERTBackbone=ResNet18, epsilon=2/2552023.03 | 32.15 | — | |
| ABSCERTBackbone=ResNet18, epsilon=4/2552023.03 | 32.25 | 212.6 | |
| ABSCERTBackbone=ResNet18, epsilon=8/2552023.03 | 32.86 | — | |
| ABSCERTBackbone=VGG11, epsilon=2/2552023.03 | 36.29 | — | |
| ABSCERTBackbone=VGG11, epsilon=4/2552023.03 | 36.35 | 2,530.3 | |
| ABSCERTBackbone=VGG11, epsilon=8/2552023.03 | 36.93 | — | |
| ABSCERTBackbone=Inception V1, epsilon=2/2552023.03 | 41.67 | — | |
| ABSCERTBackbone=Inception V1, epsilon=4/2552023.03 | 41.67 | 2,183.7 | |
| ABSCERTBackbone=Inception V1, epsilon=8/2552023.03 | 43.12 | — | |
| ABSCERTBackbone=AlexNet, epsilon=2/2552023.03 | 44.96 | — | |
| ABSCERTBackbone=AlexNet, epsilon=4/2552023.03 | 44.96 | 508.2 | |
| ABSCERTBackbone=AlexNet, epsilon=8/2552023.03 | 44.97 | — |