Class Incremental Learning on CUB200 Inc10 (test)
92.51Average AccuracyCoRe
Evaluation results
| Method | Links | ||
|---|---|---|---|
| CoReBackbone=ViT-B/16-IN21K2026.03 | 92.51 | 86.9 | |
| EASEBackbone=ViT-B/16, Pre-trained Model=ImageNet-21K2026.03 | 92.51 | 86.49 | |
| AdapterBackbone=ViT-B/16-IN21K2026.03 | 92.21 | 86.73 | |
| ViTBackbone=ViT-B/16-IN21K2026.03 | 92.2 | 86.73 | |
| FinetuneBackbone=ViT-B/16-IN21K2026.03 | 91.82 | 86.39 | |
| SSFBackbone=ViT-B/16-IN21K2026.03 | 91.72 | 86.13 | |
| APER-AdapterBackbone=ViT-B/16, Pre-trained Model=ImageNet-21K2026.03 | 91.62 | 86.72 | |
| SimpleCILBackbone=ViT-B/16, Pre-trained Model=ImageNet-21K2026.03 | 91.58 | 86.73 | |
| PromptBackbone=ViT-B/16-IN21K2026.03 | 91.02 | 84.99 | |
| PAM (RN152)Backbone=ResNet-1522026.03 | 89.91 | 88.35 | |
| PAM (RN101)Backbone=ResNet-1012026.03 | 89.76 | 87.26 | |
| PAM (RN18)Backbone=ResNet-182026.03 | 87.4 | 83.69 | |
| PAM (RN50)Backbone=ResNet-502026.03 | 85.4 | 82.67 | |
| DualPromptBackbone=ViT-B/16, Pre-trained Model=ImageNet-21K2026.03 | 81.36 | 70.51 | |
| CODA-PromptBackbone=ViT-B/16, Pre-trained Model=ImageNet-21K2026.03 | 77.65 | 68.44 | |
| L2PBackbone=ViT-B/16, Pre-trained Model=ImageNet-21K2026.03 | 73.22 | 61.55 | |
| FinetuneBackbone=ViT-B/16, Pre-trained Model=ImageNet-21K2026.03 | 55.78 | 33.13 |