Continual Learning on CIFAR-100 50 tasks
89.12Final AccuracySR2-LoRA
Evaluation results
| Method | Links | ||
|---|---|---|---|
| SR2-LoRABackbone=ViT-B/16-IN21K2026.05 | 89.12 | 93.21 | |
| SLCABackbone=ViT-B/16-IN21K2026.05 | 87.9 | 91.96 | |
| SSIATBackbone=ViT-B/16-IN21K2026.05 | 87.33 | 91.63 | |
| MOSBackbone=ViT-B/16-IN21K2026.05 | 87.05 | 91.44 | |
| LoRA-DRSBackbone=ViT-B/16-IN21K2026.05 | 86.51 | 91.17 | |
| MACILBackbone=ViT-B/16-IN21K2026.05 | 85.09 | 90.89 | |
| EASEBackbone=ViT-B/16-IN21K2026.05 | 77.43 | 84.59 | |
| CL-LoRABackbone=ViT-B/16-IN21K2026.05 | 71.14 | 78.8 | |
| CODA-PromptBackbone=ViT-B/16-IN21K2026.05 | 55.45 | 68.39 | |
| InfLoRABackbone=ViT-B/16-IN21K2026.05 | 55.19 | 69.96 |