Continual Learning on ImageNet-R 10 sequential tasks 200 classes
82.92A1:NCPrompt
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| CPrompt2024.03 | 82.92 | — | 77.14 | 5.97 | |
| ESN2024.03 | 81.63 | — | 75.11 | 5.68 | |
| DualPrompt2024.03 | 81.39 | — | 74.87 | 7.18 | |
| CODAprompt2024.03 | 81.32 | — | 75.51 | 5.91 | |
| L2P2024.03 | 80.83 | — | 74.6 | 4.52 | |
| FTP + EWCfine-tune=attention blocks2023.10 | 77.26 | 1.48 | — | — | |
| L22023.10 | 76.06 | 1.68 | — | — | |
| FTPre-initialize=after each task, fine-tune=attention blocks2023.10 | 76.06 | 2.27 | — | — | |
| CODA-P2023.10 | 75.45 | 1.64 | — | — | |
| L2P++2023.10 | 71.66 | 1.78 | — | — | |
| DualPrompt2023.10 | 71.32 | 1.71 | — | — | |
| LwF.MC2023.10 | 66.73 | 3.52 | — | — | |
| EWC2023.10 | 64.66 | 1.55 | — | — | |
| FT++2023.10 | 48.93 | 9.81 | — | — | |
| UB2024.03 | — | — | 80.27 | — |