Long-tailed Image Classification on CIFAR-100-LT λ=10 (test)
37.62Error RateRISDA
Evaluation results
| Method | Links | |
|---|---|---|
| RISDABackbone=ResNet-322021.12 | 37.62 | |
| Hybrid-PSCBackbone=ResNet-32, Reference (Source of result)=Original paper2021.12 | 37.63 | |
| MetaSAug with cross-entropy lossBackbone=ResNet-32, Reference (Source of result)=Original paper2021.12 | 38.27 | |
| Meta-class-weight with cross-entropy lossBackbone=ResNet-32, Reference (Source of result)=Li et al. 20212021.12 | 40.42 | |
| BBNBackbone=ResNet-32, Reference (Source of result)=Original paper2021.12 | 40.88 | |
| Meta-weight netBackbone=ResNet-32, Reference (Source of result)=Li et al. 20212021.12 | 41.09 | |
| MixupBackbone=ResNet-32, Reference (Source of result)=Li et al. 20212021.12 | 41.98 | |
| Class-balanced CEBackbone=ResNet-32, Reference (Source of result)=Li et al. 20212021.12 | 42.43 | |
| Class-balanced fine-tuningBackbone=ResNet-32, Reference (Source of result)=Li et al. 20212021.12 | 42.43 | |
| Cross-EntropyBackbone=ResNet-322021.12 | 44.27 | |
| L2RWBackbone=ResNet-32, Reference (Source of result)=Li et al. 20212021.12 | 47.88 |