Class-Incremental Learning on ObjectNet 20 steps (incremental)
68.12Average AccuracyBoosting-CL
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Boosting-CLBackbone=CLIP ViT-B/162026.03 | 68.12 | 55.3 | |
| SimEBackbone=CLIP ViT-B/162026.03 | 67.51 | 55.74 | |
| APERBackbone=CLIP ViT-B/162026.03 | 67.18 | 55.24 | |
| CODA-PromptBackbone=CLIP ViT-B/162026.03 | 66.07 | 53.29 | |
| MISABackbone=CLIP ViT-B/162026.03 | 64.71 | 54.25 | |
| L2PBackbone=CLIP ViT-B/162026.03 | 63.78 | 52.19 | |
| DualPromptBackbone=CLIP ViT-B/162026.03 | 59.27 | 49.33 |