Image Classification on CIFAR-10 (test) (Accuracy Drop)
1.42Accuracy DropProAct
Evaluation results
| Method | Links | |
|---|---|---|
| ProActBackbone=ResNet-50, BER=1E-62024.06 | 1.42 | |
| FitActBackbone=ResNet-50, BER=1E-62024.06 | 1.53 | |
| ProActBackbone=VGG-16, BER=3E-62024.06 | 1.88 | |
| ProActBackbone=AlexNet, BER=1E-62024.06 | 2.52 | |
| FitActBackbone=VGG-16, BER=3E-62024.06 | 2.61 | |
| FT-ClipActBackbone=VGG-16, BER=3E-62024.06 | 3.19 | |
| FitActBackbone=AlexNet, BER=1E-62024.06 | 4.34 | |
| FT-ClipActBackbone=AlexNet, BER=1E-62024.06 | 7.67 | |
| FT-ClipActBackbone=ResNet-50, BER=1E-62024.06 | 9.09 |