Long-tailed Image Classification on CIFAR-100 long-tailed (IF=100) (test)
57.7Top-1 AccuracyGLMC + SAMN
Evaluation results
| Method | Links | |
|---|---|---|
| GLMC + SAMNBackbone=ResNet32, Imbalance Factor=1002026.06 | 57.7 | |
| GLMCBackbone=ResNet32, Imbalance Factor=1002026.06 | 56.8 | |
| SLAS + SAMNBackbone=ResNet32, Imbalance Factor=1002026.06 | 54.1 | |
| CE + SAMNBackbone=ResNet32, Imbalance Factor=1002026.06 | 54 | |
| WD + MaxNormBackbone=ResNet32, Imbalance Factor=1002026.06 | 53.4 | |
| IWBBackbone=ResNet32, Imbalance Factor=1002026.06 | 53.3 | |
| SLASBackbone=ResNet32, Imbalance Factor=1002026.06 | 52.7 | |
| DiffuLTBackbone=ResNet32, Imbalance Factor=1002026.06 | 51.5 | |
| RIDE (3 experts)Backbone=ResNet32, Imbalance Factor=1002026.06 | 48 | |
| CUDABackbone=ResNet32, Imbalance Factor=1002026.06 | 47.6 | |
| CE (with WD in stage one)Backbone=ResNet32, Imbalance Factor=1002026.06 | 47.4 | |
| CMOBackbone=ResNet32, Imbalance Factor=1002026.06 | 47.2 | |
| CSABackbone=ResNet32, Imbalance Factor=1002026.06 | 46.6 | |
| ADRWBackbone=ResNet32, Imbalance Factor=1002026.06 | 46.4 | |
| DiVEBackbone=ResNet32, Imbalance Factor=1002026.06 | 45.4 | |
| SAMBackbone=ResNet32, Imbalance Factor=1002026.06 | 45.4 | |
| cRTBackbone=ResNet32, Imbalance Factor=1002026.06 | 45.3 | |
| FBLBackbone Net=ResNet-32, IF=1002023.05 | 45.22 | |
| LABackbone Net=ResNet-32, IF=1002023.05 | 43.89 | |
| BBNBackbone=ResNet32, Imbalance Factor=1002026.06 | 42.6 | |
| BBNBackbone Net=ResNet-32, IF=1002023.05 | 42.56 | |
| LDAM-DRWBackbone Net=ResNet-32, IF=1002023.05 | 42.04 | |
| LDAM-DRWBackbone=ResNet32, Imbalance Factor=1002026.06 | 42 | |
| CAM-BSBackbone=ResNet32, Imbalance Factor=1002026.06 | 41.7 | |
| CE lossBackbone Net=ResNet-32, IF=1002023.05 | 39.43 | |
| Focal LossBackbone=ResNet32, Imbalance Factor=1002026.06 | 38.4 |