Lifelong Learning on Split miniImageNet (test)
69.46AccuracyMultitask
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| MultitaskNetwork=multi-head2022.11 | 69.46 | — | — | |
| CUBERNetwork=multi-head2022.11 | 62.67 | — | 0.23 | |
| TRGPNetwork=multi-head2022.11 | 61.78 | — | -0.5 | |
| GPMNetwork=multi-head2022.11 | 60.41 | — | -0.7 | |
| HATNetwork=multi-head2022.11 | 60.23 | — | -4 | |
| ER_ResNetwork=multi-head2022.11 | 58.04 | — | -7 | |
| TAG-RMSProp2021.05 | 57.2 | -5.52 | — | |
| A-GEMNetwork=multi-head2022.11 | 56.88 | — | -13 | |
| ER2021.05 | 54.67 | -9.39 | — | |
| Stable SGD2021.05 | 51.81 | -8.18 | — | |
| EWCNetwork=multi-head2022.11 | 50.78 | — | -12 | |
| A-GEM2021.05 | 50.32 | -10.69 | — | |
| Naive SGD2021.05 | 48.19 | -13.83 | — | |
| EWC2021.05 | 47.87 | -13.79 | — | |
| Naive RMSProp2021.05 | 45.06 | -19.32 | — |