Continual Learning on ImageNet sub
77.24AccuracyAdaB2N
Evaluation Results
| Method | Links | |
|---|---|---|
| AdaB2NEvaluation Protocol=Task-IL, Incremental Phases=10, Experience Replay (ER)=true, Batch Size (|B|)=32, Memory Size (|M|)=2000, Input Size=224x2242023.10 | 77.24 | |
| CNEvaluation Protocol=Task-IL, Incremental Phases=10, Experience Replay (ER)=true, Batch Size (|B|)=32, Memory Size (|M|)=2000, Input Size=224x2242023.10 | 76.48 | |
| BNEvaluation Protocol=Task-IL, Incremental Phases=10, Experience Replay (ER)=true, Batch Size (|B|)=32, Memory Size (|M|)=2000, Input Size=224x2242023.10 | 72.92 | |
| AdaB2NEvaluation Protocol=Class-IL, Incremental Phases=10, Experience Replay (ER)=true, Batch Size (|B|)=32, Memory Size (|M|)=2000, Input Size=224x2242023.10 | 22.08 | |
| CNEvaluation Protocol=Class-IL, Incremental Phases=10, Experience Replay (ER)=true, Batch Size (|B|)=32, Memory Size (|M|)=2000, Input Size=224x2242023.10 | 21.88 | |
| BNEvaluation Protocol=Class-IL, Incremental Phases=10, Experience Replay (ER)=true, Batch Size (|B|)=32, Memory Size (|M|)=2000, Input Size=224x2242023.10 | 16.86 |