Machine Unlearning on Tiny-ImageNet Forget 1%
61.22Test AccuracyNEGGRAD
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| NEGGRADBackbone=Swin-T2026.03 | 61.22 | 87.78 | 19.2 | |
| BASEBackbone=Swin-T2026.03 | 61.21 | 87.77 | 19.2 | |
| RETRAINBackbone=Swin-T2026.03 | 60.9 | 49.81 | 0 | |
| SALUNBackbone=Swin-T2026.03 | 53.16 | 46.36 | 3.73 | |
| AMUNBackbone=Swin-T2026.03 | 52.59 | 51.03 | 6.55 | |
| FINETUNEBackbone=Swin-T2026.03 | 52.31 | 62.48 | 6.6 | |
| REGUNBackbone=Swin-T2026.03 | 52.26 | 45.07 | 5.82 | |
| ℓ1-SPARSEBackbone=Swin-T2026.03 | 51.94 | 62.37 | 6.54 | |
| NEGGRAD+Backbone=Swin-T2026.03 | 48.99 | 66.99 | 10.52 | |
| SSDBackbone=Swin-T2026.03 | 41.99 | 68.08 | 17.63 |