Class-wise Federated Unlearning on Caltech-101
99.93Retain AccuracyFT
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| FTModel=ViT, Number of Clients=10, Data Distribution=IID2025.03 | 99.93 | 77.39 | 46.69 | 71.53 | 24.68 | 1,190,000,000,000 | 6,600,000,000,000,000,000,000 | |
| NoTModel=ViT, Number of Clients=10, Data Distribution=IID2025.03 | 99.36 | 0.7 | 46.54 | 84.83 | 2.07 | 890,000,000,000 | 4,950,000,000,000,000,000,000 | |
| RetrainModel=ViT, Number of Clients=10, Data Distribution=IID2025.03 | 99.32 | 0 | 46.85 | 92.07 | 0 | 1,480,000,000,000 | 8,240,000,000,000,000,000,000 | |
| MODEModel=ViT, Number of Clients=10, Data Distribution=IID2025.03 | 98.74 | 0 | 46.97 | 84.7 | 2.28 | 1,350,000,000,000 | 7,540,000,000,000,000,000,000 |