Image Classification on CIFAR-LT-100 (Imbalance Factor 100, test)
51.99Top-1 ErrorMetaSAug with LDAM loss
Evaluation results
| Method | Links | |
|---|---|---|
| MetaSAug with LDAM lossBackbone=ResNet-32, Imbalance factor=1002021.03 | 51.99 | |
| MetaSAug with cross-entropy lossBackbone=ResNet-32, Imbalance factor=1002021.03 | 53.13 | |
| MetaSAug with focal lossBackbone=ResNet-32, Imbalance factor=1002021.03 | 54.11 | |
| Meta-class-weight with focal lossBackbone=ResNet-32, Imbalance factor=1002021.03 | 55.3 | |
| Meta-class-weight with LDAM lossBackbone=ResNet-32, Imbalance factor=1002021.03 | 55.92 | |
| Meta-class-weight with cross-entropy lossBackbone=ResNet-32, Imbalance factor=1002021.03 | 56.65 | |
| LDAM-DRWBackbone=ResNet-32, Imbalance factor=1002021.03 | 57.11 | |
| BBNBackbone=ResNet-32, Imbalance factor=1002021.03 | 57.44 | |
| Meta-weight netBackbone=ResNet-32, Imbalance factor=1002021.03 | 58.39 | |
| Class-balanced fine-tuningBackbone=ResNet-32, Imbalance factor=1002021.03 | 58.5 | |
| LDAM lossBackbone=ResNet-32, Imbalance factor=1002021.03 | 59.4 | |
| Class-balanced focal lossBackbone=ResNet-32, Imbalance factor=1002021.03 | 60.4 | |
| MixupBackbone=ResNet-32, Imbalance factor=1002021.03 | 60.46 | |
| L2RWBackbone=ResNet-32, Imbalance factor=1002021.03 | 61.1 | |
| Class-balanced cross-entropy lossBackbone=ResNet-32, Imbalance factor=1002021.03 | 61.23 | |
| Cross-entropy trainingBackbone=ResNet-32, Imbalance factor=1002021.03 | 61.54 | |
| Focal lossBackbone=ResNet-32, Imbalance factor=1002021.03 | 61.59 |