Task-Incremental Learning on Stanford Cars
79.92Average AccuracyPrompt
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| PromptBackbone=ViT-B/16-IN21K2026.03 | 79.92 | — | — | — | 80.74 | |
| CoReBackbone=ViT-B/16-IN21K2026.03 | 79.41 | — | — | — | 80.4 | |
| SSFBackbone=ViT-B/16-IN21K2026.03 | 75.11 | — | — | — | 75.46 | |
| AdapterBackbone=ViT-B/16-IN21K2026.03 | 72.22 | — | — | — | 73.05 | |
| ViTBackbone=ViT-B/16-IN21K2026.03 | 72.15 | — | — | — | 72.89 | |
| FinetuneBackbone=ViT-B/16-IN21K2026.03 | 70.66 | — | — | — | 69.63 | |
| CIP-Net2025.12 | — | 86.2 | 92.1 | 92.5 | — | |
| EWC2025.12 | — | 45.6 | 31.5 | 28.7 | — | |
| ICICLE2025.12 | — | 65.4 | 64.5 | 58.3 | — | |
| LWF2025.12 | — | 37.5 | 35.6 | 25 | — | |
| LWM2025.12 | — | 45.9 | 41.6 | 30.5 | — |