Long-tailed Image Classification on CIFAR100-LT imbalance factor 100 (test)
77.8Top-1 AccBALLAD
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| BALLADBackbone=ViT-B/162021.11 | 77.8 | 84.9 | 79.7 | 67.3 | |
| BALLADBackbone=ResNet-50x162021.11 | 63.7 | 74.6 | 62.8 | 52 | |
| BALLADBackbone=ResNet-1012021.11 | 59.2 | 69.5 | 59.3 | 47.1 | |
| BALLADBackbone=ResNet-502021.11 | 51.6 | 62.4 | 52.3 | 38.2 | |
| TADEBackbone=ResNet-322021.11 | 49.8 | 65.4 | 49.3 | 29.3 | |
| ACENumber of experts=42021.08 | 49.6 | — | — | — | |
| ACENumber of experts=32021.08 | 49.4 | — | — | — | |
| RIDE2021.08 | 49.1 | — | — | — | |
| RIDEBackbone=ResNet-322021.11 | 49.1 | 69.3 | 49.3 | 26 | |
| CAM2021.08 | 47.8 | — | — | — | |
| Logit Adj.2021.08 | 43.9 | — | — | — | |
| LDAM+M2m2021.08 | 43.5 | — | — | — | |
| CRTBackbone=ResNet-322021.11 | 43.3 | 64 | 44.8 | 18.1 | |
| Tau-normalizedBackbone=ResNet-322021.11 | 43.2 | 65.7 | 43.6 | 17.3 | |
| LFME2021.08 | 42.3 | — | — | — | |
| LDAM+DRW2021.08 | 42 | — | — | — | |
| LDAM+DRWBackbone=ResNet-322021.11 | 42 | 61.5 | 41.7 | 20.2 | |
| OLTRBackbone=ResNet-322021.11 | 41.2 | 61.8 | 41.4 | 17.6 | |
| Mixup2021.08 | 39.5 | — | — | — | |
| BBN2021.08 | 39.4 | — | — | — | |
| CB Focal loss2021.08 | 38.7 | — | — | — | |
| Baseline2021.08 | 38.3 | — | — | — | |
| Focal loss2021.08 | 37.4 | — | — | — |