Machine Unlearning on CIFAR-100 Forget 1%
75.52Test AccuracyBASE
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| BASEBackbone=ResNet-182026.03 | 75.52 | 76.09 | 12.84 | |
| RETRAINBackbone=ResNet-182026.03 | 75.33 | 49.56 | 0 | |
| SSDBackbone=ResNet-182026.03 | 75.12 | 75.78 | 12.63 | |
| NEGGRADBackbone=ResNet-182026.03 | 72.35 | 73.79 | 12.26 | |
| AMUNBackbone=ResNet-182026.03 | 70.43 | 57.91 | 5.96 | |
| REGUNBackbone=ResNet-182026.03 | 69.23 | 46.91 | 3.01 | |
| NEGGRAD+Backbone=ResNet-182026.03 | 69.1 | 69.96 | 11.13 | |
| SALUNBackbone=ResNet-182026.03 | 67.33 | 51.18 | 3.29 | |
| FINETUNEBackbone=ResNet-182026.03 | 66.67 | 59.76 | 7.04 | |
| ℓ1-SPARSEBackbone=ResNet-182026.03 | 66.3 | 59.78 | 7.02 |