Long-tail Class Incremental Learning on ImageNet-LT 20 tasks 1.0
75.6AccuracyUpper Bound
Evaluation Results
| Method | Links | |
|---|---|---|
| Upper BoundBackbone=ViT-B/16, Evaluation Protocol=Joint Learning2026.03 | 75.6 | |
| OursBackbone=ViT-B/162026.03 | 68.2 | |
| CLIP Zero-ShotBackbone=ViT-B/16, Evaluation Protocol=Zero-Shot2026.03 | 63.1 | |
| Adapter Fine-tuningBackbone=ViT-B/162026.03 | 19.6 |