Image Classification on CIFAR-100 Long-tailed Imbalance 200 (train)
37.91Top-1 AccuracyMeta-Weight Net
Evaluation results
| Method | Links | |
|---|---|---|
| Meta-Weight NetBackbone=ResNet-322020.10 | 37.91 | |
| CDB-CEBackbone=ResNet-32, tau=dynamically updated2020.10 | 37.4 | |
| Equalization lossBackbone=ResNet-32, Data Augmentation=Without additional augmentation2020.10 | 37.34 | |
| Class-BalancedBackbone=ResNet-322020.10 | 36.23 | |
| Focal lossBackbone=ResNet-322020.10 | 35.62 | |
| L2RWBackbone=ResNet-322020.10 | 33.38 |