Image Classification on CIFAR-100 (test) (Top-1 Accuracy Change)
11.97Top-1 Accuracy ChangeNC-CTD
Evaluation results
| Method | Links | |
|---|---|---|
| NC-CTDBackbone=ResNet-34, Compression Rate=0.32026.05 | 11.97 | |
| APSSFBackbone=ResNet-34, Compression Rate=0.52026.05 | 0.003 | |
| Slice-wise Feature DistillationBackbone=ResNet-34, Compression Rate=0.52026.05 | -0.98 | |
| LJSVDBackbone=ResNet-34, Compression Rate=0.952026.05 | -1.42 | |
| EdropoutBackbone=ResNet-34, Compression Rate=0.52026.05 | -5.17 | |
| Slice-wise Feature DistillationBackbone=ResNet-34, Compression Rate=0.72026.05 | -5.57 |