Class-Incremental Learning on TinyImageNet 5 Steps (test)
86.47Average AccuracySimE
Evaluation results
| Method | Links | ||
|---|---|---|---|
| SimEBackbone=CLIP ViT-L/14, Pre-trained Dataset=Laion-2B2026.03 | 86.47 | 83.33 | |
| ZSCL2026.03 | 80.27 | 73.57 | |
| SimE2026.03 | 79.35 | 75.37 | |
| LwF-VR2026.03 | 77.56 | 70.89 | |
| iCaRL2026.03 | 77.02 | 70.39 | |
| Continual-CLIP2026.03 | 70.49 | 66.43 | |
| CLIP2026.03 | 69.62 | 65.3 | |
| Fine-tune2026.03 | 61.54 | 46.66 | |
| LwF2026.03 | 60.97 | 48.77 | |
| DyTox2026.03 | 55.58 | 47.23 | |
| UCIR2026.03 | 50.3 | 39.42 | |
| PASS2026.03 | 49.54 | 41.64 |