Class-incremental Learning on CIFAR100 (B=2, C=2) (test)
78Average AccuracyOracle w/ LingoCL
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Oracle w/ LingoCLBackbone=ResNet-182024.03 | 78 | 78 | — | |
| OracleBackbone=ResNet-182024.03 | 77.6 | 77.6 | — | |
| DyTox w/ LingoCLBackbone=ConViT2024.03 | 65.9 | 46.3 | 36.1 | |
| DyToxBackbone=ConViT2024.03 | 64.5 | 44.8 | 41.3 | |
| AANet w/ LingoCLBackbone=ResNet-182024.03 | 58.7 | 38.6 | — | |
| AANetBackbone=ResNet-182024.03 | 57.7 | 37.6 | — | |
| IL2M w/ LingoCLBackbone=ResNet-182024.03 | 48 | 39.6 | 42.8 | |
| LUCIR w/ LingoCLBackbone=ResNet-182024.03 | 46.8 | 37 | 42.3 | |
| CwD w/ LingoCLBackbone=ResNet-182024.03 | 46 | 38.4 | 36.9 | |
| LUCIRBackbone=ResNet-182024.03 | 45.6 | 36.2 | 44.5 | |
| BiC w/ LingoCLBackbone=ResNet-182024.03 | 44.9 | 31.1 | 29.7 | |
| IL2MBackbone=ResNet-182024.03 | 44 | 34.2 | 48.5 | |
| BiMeCo w/ LingoCLBackbone=MobileNet-V2†2024.03 | 43.4 | 32.5 | 45 | |
| CwDBackbone=ResNet-182024.03 | 40.2 | 34 | 44.6 | |
| BiCBackbone=ResNet-182024.03 | 38.1 | 23.6 | 38.6 | |
| BiMeCoBackbone=MobileNet-V2†2024.03 | 37.2 | 25.9 | 50.7 |