Few-Shot Class-Incremental Image-to-Text on SUN397
66.11AccuracyCoACT
Evaluation results
| Method | Links | |
|---|---|---|
| CoACTBackbone=ViT-B/162024.05 | 66.11 | |
| LoRA2024.05 | 65.9 | |
| Linear Tuning2024.05 | 65.87 | |
| Prototype Learning2024.05 | 65.68 | |
| PriViLege2024.05 | 65.09 | |
| CPE-CLIP2024.05 | 64.48 |