Class-level Machine Unlearning on ImageNet (test)
54.18Retained Accuracy (Test)Original Model
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Original ModelBackbone=ViT-B/32, Architecture=CLIP2025.06 | 54.18 | 55.75 | — | |
| SpotterBackbone=ViT-B/32, Architecture=CLIP2025.06 | 0 | 55 | 0.0141 |