Image Classification on CIFAR-10 Step-Imbalance
88.01AccuracyIB + CB
Evaluation results
| Method | Links | |
|---|---|---|
| IB + CBBackbone=ResNet-32, Imbalance Ratio (p)=102021.10 | 88.01 | |
| IB + FocalBackbone=ResNet-32, Imbalance Ratio (p)=102021.10 | 87.9 | |
| LDAM-DRWBackbone=ResNet-32, Imbalance Ratio (p)=102021.10 | 87.81 | |
| IBBackbone=ResNet-32, Imbalance Ratio (p)=102021.10 | 87.72 | |
| IB + FocalBackbone=ResNet-32, Imbalance Ratio (p)=202021.10 | 85.68 | |
| IBBackbone=ResNet-32, Imbalance Ratio (p)=202021.10 | 85.41 | |
| IB + CBBackbone=ResNet-32, Imbalance Ratio (p)=202021.10 | 85.27 | |
| LDAMBackbone=ResNet-32, Imbalance Ratio (p)=102021.10 | 85 | |
| Baseline (CE)Backbone=ResNet-32, Imbalance Ratio (p)=102021.10 | 84.16 | |
| IB + FocalBackbone=ResNet-32, Imbalance Ratio (p)=502021.10 | 82.38 | |
| IB + CBBackbone=ResNet-32, Imbalance Ratio (p)=502021.10 | 82.09 | |
| IBBackbone=ResNet-32, Imbalance Ratio (p)=502021.10 | 81.66 | |
| Baseline (CE)Backbone=ResNet-32, Imbalance Ratio (p)=202021.10 | 79.71 | |
| IB + FocalBackbone=ResNet-32, Imbalance Ratio (p)=1002021.10 | 77.97 | |
| LDAM-DRWBackbone=ResNet-32, Imbalance Ratio (p)=1002021.10 | 76.92 | |
| IBBackbone=ResNet-32, Imbalance Ratio (p)=1002021.10 | 76.53 | |
| IB + CBBackbone=ResNet-32, Imbalance Ratio (p)=1002021.10 | 75.97 | |
| IB + FocalBackbone=ResNet-32, Imbalance Ratio (p)=2002021.10 | 74.12 | |
| IBBackbone=ResNet-32, Imbalance Ratio (p)=2002021.10 | 72.15 | |
| IB + CBBackbone=ResNet-32, Imbalance Ratio (p)=2002021.10 | 69.96 | |
| Baseline (CE)Backbone=ResNet-32, Imbalance Ratio (p)=502021.10 | 69.35 | |
| LDAMBackbone=ResNet-32, Imbalance Ratio (p)=2002021.10 | 66.58 | |
| Baseline (CE)Backbone=ResNet-32, Imbalance Ratio (p)=1002021.10 | 64.81 | |
| Baseline (CE)Backbone=ResNet-32, Imbalance Ratio (p)=2002021.10 | 56.97 |