Image Classification on Long-tailed CIFAR-100 Imbalance factor 0.1 (test)
63.73Top-1 AccResLT (3 experts)
Evaluation results
| Method | Links | |
|---|---|---|
| ResLT (3 experts)One-stage=false, Backbone=ResNet-32, experts=32021.01 | 63.73 | |
| ResLTSOne-stage=true, Backbone=ResNet-32, Strong data augmentation=true2021.01 | 62.01 | |
| ResLTOne-stage=true, Backbone=ResNet-322021.01 | 60.79 | |
| Causal NormOne-stage=true, Backbone=ResNet-322021.01 | 59.6 | |
| BBNOne-stage=true, Backbone=ResNet-322021.01 | 59.12 | |
| ELF(LDAM)+DRWOne-stage=false, Backbone=ResNet-322021.01 | 58.9 | |
| LDAM+DRWOne-stage=false, Backbone=ResNet-322021.01 | 58.71 | |
| CE+DRWOne-stage=false, Backbone=ResNet-322021.01 | 58.12 | |
| CE+DRSOne-stage=false, Backbone=ResNet-322021.01 | 58.11 | |
| MixupOne-stage=true, Backbone=ResNet-32, Strong data augmentation=true2021.01 | 58.02 | |
| CB-FocalOne-stage=true, Backbone=ResNet-32, Strong data augmentation=true2021.01 | 57.99 | |
| LDAMOne-stage=true, Backbone=ResNet-322021.01 | 56.91 | |
| FocalOne-stage=true, Backbone=ResNet-32, Strong data augmentation=true2021.01 | 55.78 |