Long-tailed Image Classification on CIFAR-10 LT Imbalance ratio 50
85.46Top-1 AccGCL
Evaluation results
| Method | Links | |
|---|---|---|
| GCLBackbone Net=ResNet-322023.05 | 85.46 | |
| Contrastive learningBackbone Net=ResNet-322023.05 | 85.36 | |
| MisLASBackbone Net=ResNet-322023.05 | 85.16 | |
| CE loss + mixup + cRTBackbone Net=ResNet-322023.05 | 84.21 | |
| De-confound-TDEBackbone Net=ResNet-322023.05 | 83.6 | |
| BBNBackbone Net=ResNet-322023.05 | 81.18 | |
| LDAM-DRWBackbone Net=ResNet-322023.05 | 81.03 | |
| CE loss + mixupBackbone Net=ResNet-322023.05 | 79.48 | |
| CE lossBackbone Net=ResNet-322023.05 | 74.81 |