Class-Incremental Learning on CIFAR100 10 steps
85.94Average AccuracySimE
Evaluation results
| Method | Links | ||
|---|---|---|---|
| SimEBackbone=CLIP ViT-B/162026.03 | 85.94 | 77.1 | |
| Boosting-CLBackbone=CLIP ViT-B/162026.03 | 85.21 | 77.52 | |
| L2PBackbone=CLIP ViT-B/162026.03 | 81.9 | 73.08 | |
| DualPromptBackbone=CLIP ViT-B/162026.03 | 81.45 | 72.51 | |
| MISABackbone=CLIP ViT-B/162026.03 | 81.2 | 80.63 | |
| CODA-PromptBackbone=CLIP ViT-B/162026.03 | 76.98 | 62.25 | |
| APERBackbone=CLIP ViT-B/162026.03 | 75.76 | 65.5 |