Long-tailed Image Classification on CIFAR-100-LT λ=200 (test)
55.24Error RateRISDA
Evaluation results
| Method | Links | |
|---|---|---|
| RISDABackbone=ResNet-322021.12 | 55.24 | |
| MetaSAug with cross-entropy lossBackbone=ResNet-32, Reference (Source of result)=Original paper2021.12 | 60.06 | |
| Meta-class-weight with cross-entropy lossBackbone=ResNet-32, Reference (Source of result)=Li et al. 20212021.12 | 60.69 | |
| Class-balanced fine-tuningBackbone=ResNet-32, Reference (Source of result)=Li et al. 20212021.12 | 61.34 | |
| Meta-weight netBackbone=ResNet-32, Reference (Source of result)=Li et al. 20212021.12 | 63.38 | |
| Class-balanced CEBackbone=ResNet-32, Reference (Source of result)=Li et al. 20212021.12 | 64.44 | |
| Cross-EntropyBackbone=ResNet-322021.12 | 65.3 | |
| L2RWBackbone=ResNet-32, Reference (Source of result)=Li et al. 20212021.12 | 67 |