Continual Learning on ImageNet-R (20 tasks)
86.26Average Accuracy (20 Tasks)LoDA + CA
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| LoDA + CAFeature Replay=true2026.02 | 86.26 | — | 81.13 | — | — | |
| LoDAFeature Replay=false2026.02 | 84.94 | — | 78.96 | — | — | |
| MACILVenue=ICML-25, Feature Replay=true2026.02 | 84.34 | — | 79.43 | — | — | |
| SR2-LoRABackbone=ViT-B/16-IN21K, Number of Tasks=202026.05 | 84.32 | — | 79.54 | — | — | |
| MACILBackbone=ViT-B/16-IN21K, Number of Tasks=202026.05 | 84.25 | — | 79.46 | — | — | |
| CoSOVenue=NIPS-25, Feature Replay=false2026.02 | 83.69 | — | 78.19 | — | — | |
| CL-LoRABackbone=ViT-B/16-IN21K, Number of Tasks=202026.05 | 83.45 | — | 77.2 | — | — | |
| LITELORATask Order=Order 12026.06 | 82.83 | — | — | 7.84 | 6 | |
| LoRA-P&MVenue=NIPS-25, Feature Replay=false2026.02 | 82.77 | — | 76.37 | — | — | |
| SD-LoRATask Order=Order 12026.06 | 82.41 | — | — | 10.25 | 20 | |
| SSIATBackbone=ViT-B/16-IN21K, Number of Tasks=202026.05 | 82.35 | — | 78.31 | — | — | |
| SSIATVenue=CVPR-24, Feature Replay=true2026.02 | 82.3 | — | 75.67 | — | — | |
| LITELORATask Order=Order 32026.06 | 81.71 | — | — | 12.26 | 6 | |
| LITELORATask Order=Order 22026.06 | 81.64 | — | — | 8.49 | 7 | |
| SD-LoRATask Order=Order 22026.06 | 81.55 | — | — | 8.89 | 20 | |
| SD-LoRATask Order=Order 32026.06 | 81.27 | — | — | 13.67 | 20 | |
| jointBackbone=ViT-B/16, Pre-training=ImageNet-21K2024.03 | 81.14 | 81.14 | — | — | — | |
| LoRA-DRSBackbone=ViT-B/16-IN21K, Number of Tasks=202026.05 | 80.66 | — | 74.96 | — | — | |
| SLCABackbone=ViT-B/16-IN21K, Number of Tasks=202026.05 | 80.65 | — | 75.53 | — | — | |
| MOSBackbone=ViT-B/16-IN21K, Number of Tasks=202026.05 | 80.39 | — | 75.04 | — | — | |
| EASEBackbone=ViT-B/16-IN21K, Number of Tasks=202026.05 | 80.28 | — | 73.78 | — | — | |
| SD-LoRAVenue=ICLR-25, Feature Replay=false2026.02 | 80.22 | — | 75.26 | — | — | |
| SLCAVenue=ICCV-23, Feature Replay=true2026.02 | 79.92 | — | 74.63 | — | — | |
| InfLoRABackbone=ViT-B/16-IN21K, Number of Tasks=202026.05 | 79.76 | — | 73.01 | — | — | |
| Bi-LoRAVenue=CVPR-25, Feature Replay=false2026.02 | 79.28 | — | 72.41 | — | — | |
| InfLoRABackbone=ViT-B/16, Pre-training=ImageNet-21K, Inserted blocks=All Transformer blocks2024.03 | 77.28 | 71.01 | — | — | — | |
| InfLoRAVenue=CVPR-24, Feature Replay=false2026.02 | 77.28 | — | 71.01 | — | — | |
| CODAPromptVenue=CVPR’23, Feature Replay=false2026.02 | 75.34 | — | 69.96 | — | — | |
| CODA-PromptBackbone=ViT-B/16-IN21K, Number of Tasks=202026.05 | 75.34 | — | 69.96 | — | — | |
| L2PVenue=CVPR’22, Feature Replay=false2026.02 | 75.28 | — | 69.64 | — | — | |
| InfLoRA-b5Backbone=ViT-B/16, Pre-training=ImageNet-21K, Inserted blocks=Bottom 5 Transformer blocks2024.03 | 74 | 68.41 | — | — | — | |
| LAEBackbone=ViT-B/16, Pre-training=ImageNet-21K2024.03 | 73.72 | 66.98 | — | — | — | |
| DualPromptVenue=ECCV’22, Feature Replay=false2026.02 | 72.45 | — | 66.61 | — | — | |
| CODA-PBackbone=ViT-B/16, Pre-training=ImageNet-21K2024.03 | 72.34 | 67.28 | — | — | — | |
| C-LoRABackbone=ViT-B/16, Pre-training=ImageNet-21K2024.03 | 70.63 | 65.71 | — | — | — | |
| DualPromptBackbone=ViT-B/16, Pre-training=ImageNet-21K2024.03 | 65.8 | 61 | — | — | — | |
| L2PBackbone=ViT-B/16, Pre-training=ImageNet-21K2024.03 | 64.57 | 57.92 | — | — | — | |
| sequentialBackbone=ViT-B/16, Pre-training=ImageNet-21K2024.03 | 51.15 | 34.62 | — | — | — |