Long-tailed Image Classification on CIFAR-100-LT λ=50 (test)
46.16Error RateRISDA
Evaluation results
| Method | Links | |
|---|---|---|
| RISDABackbone=ResNet-322021.12 | 46.16 | |
| MetaSAug with cross-entropy lossBackbone=ResNet-32, Reference (Source of result)=Original paper2021.12 | 48.1 | |
| Bag of TricksBackbone=ResNet-32, Reference (Source of result)=Original paper2021.12 | 48.31 | |
| Hybrid-PSCBackbone=ResNet-32, Reference (Source of result)=Original paper2021.12 | 51.07 | |
| Meta-class-weight with cross-entropy lossBackbone=ResNet-32, Reference (Source of result)=Li et al. 20212021.12 | 51.47 | |
| BBNBackbone=ResNet-32, Reference (Source of result)=Original paper2021.12 | 52.98 | |
| Class-balanced fine-tuningBackbone=ResNet-32, Reference (Source of result)=Li et al. 20212021.12 | 53.78 | |
| Meta-weight netBackbone=ResNet-32, Reference (Source of result)=Li et al. 20212021.12 | 54.34 | |
| MixupBackbone=ResNet-32, Reference (Source of result)=Li et al. 20212021.12 | 55.01 | |
| Class-balanced CEBackbone=ResNet-32, Reference (Source of result)=Li et al. 20212021.12 | 55.21 | |
| Cross-EntropyBackbone=ResNet-322021.12 | 55.98 | |
| L2RWBackbone=ResNet-32, Reference (Source of result)=Li et al. 20212021.12 | 56.83 |