Class Incremental Learning on Stanford Dogs 10 Tasks
92.64Average AccuracyGDDSG
Evaluation results
| Method | Links | ||
|---|---|---|---|
| GDDSGBackbone=ViT-B/16, Pre-training=ImageNet-21K2025.02 | 92.64 | 1.42 | |
| RanPACBackbone=ViT-B/16, Pre-training=ImageNet-21K2025.02 | 85.37 | 5.78 | |
| ADAMBackbone=ViT-B/16, Pre-training=ImageNet-21K2025.02 | 83.98 | 6.08 | |
| SimpleCILBackbone=ViT-B/16, Pre-training=ImageNet-21K2025.02 | 83.23 | 5.86 | |
| EASEBackbone=ViT-B/16, Pre-training=ImageNet-21K2025.02 | 81.03 | 8.18 | |
| CODA-PromptBackbone=ViT-B/16, Pre-training=ImageNet-21K2025.02 | 74.09 | 10.05 | |
| DualPromptBackbone=ViT-B/16, Pre-training=ImageNet-21K2025.02 | 70.9 | 10.68 | |
| L2PBackbone=ViT-B/16, Pre-training=ImageNet-21K2025.02 | 65.09 | 9.44 | |
| FinetuneBackbone=ViT-B/16, Pre-training=ImageNet-21K2025.02 | 45.64 | 46.19 |