Image Classification on CIFAR-100 (test) (Accuracy Drop)
4.4Accuracy DropProAct
Evaluation results
| Method | Links | |
|---|---|---|
| ProActBackbone=VGG-16, BER=1E-72024.06 | 4.4 | |
| ProActBackbone=AlexNet, BER=3E-72024.06 | 5.28 | |
| FitActBackbone=AlexNet, BER=3E-72024.06 | 6.31 | |
| FitActBackbone=VGG-16, BER=1E-72024.06 | 6.34 | |
| FT-ClipActBackbone=VGG-16, BER=1E-72024.06 | 6.74 | |
| FT-ClipActBackbone=AlexNet, BER=3E-72024.06 | 7.89 | |
| ProActBackbone=ResNet-50, BER=3E-72024.06 | 9.37 | |
| FitActBackbone=ResNet-50, BER=3E-72024.06 | 11.24 | |
| FT-ClipActBackbone=ResNet-50, BER=3E-72024.06 | 12.75 |