Image Classification on CIFAR-10-LT IR=10 (test)
74.67Accuracy (Clean)TAET
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| TAETBackbone=ResNet-18, Imbalance Ratio (IR)=102025.03 | 74.67 | 39.59 | — | 30.57 | 35.15 | 34.29 | 57.59 | |
| REATBackbone=ResNet-18, Imbalance Ratio (IR)=102025.03 | 74.56 | 31.42 | — | 22.69 | 24.02 | 22.52 | 11.07 | |
| TAETModel=ResNet502025.03 | 74.48 | 40.17 | 34.7 | 31.27 | — | — | — | |
| AT-BSLModel=ResNet502025.03 | 74.34 | 35.41 | 29.01 | 27.14 | — | — | — | |
| RoBalBackbone=ResNet-18, Imbalance Ratio (IR)=102025.03 | 73.18 | 33.14 | — | 24.13 | 27.12 | 26.98 | 13.44 | |
| AT-BSLBackbone=ResNet-18, Imbalance Ratio (IR)=102025.03 | 72.74 | 34.13 | — | 25.26 | 26.86 | 25.62 | 15.67 | |
| ATBackbone=ResNet-18, Imbalance Ratio (IR)=102025.03 | 69 | 32.53 | — | 24.28 | 25.69 | 24.55 | 15.22 | |
| HEBackbone=ResNet-18, Imbalance Ratio (IR)=102025.03 | 68.18 | 32.4 | — | 24.32 | 25.66 | 24.59 | 12.94 | |
| ADTBackbone=ResNet-18, Imbalance Ratio (IR)=102025.03 | 68.08 | 32.7 | — | 25.06 | 26 | 25.27 | 10.05 | |
| TRADESBackbone=ResNet-18, Imbalance Ratio (IR)=102025.03 | 65.77 | 25.73 | — | 19.63 | 20.23 | 19.59 | 27.42 | |
| LAS-ATBackbone=ResNet-18, Imbalance Ratio (IR)=102025.03 | 64.04 | 33.04 | — | 24.71 | 27.8 | 26.74 | 36.77 | |
| GAIRATBackbone=ResNet-18, Imbalance Ratio (IR)=102025.03 | 61.07 | 28.9 | — | 24.59 | 24.79 | 23.81 | 42.62 | |
| MARTBackbone=ResNet-18, Imbalance Ratio (IR)=102025.03 | 58.33 | 33.67 | — | 26.04 | 30.1 | 29.36 | 48.12 | |
| TAETModel=ViT-B/162025.03 | 22.65 | 18.45 | 18.56 | 14.72 | — | — | — | |
| AT-BSLModel=ViT-B/162025.03 | 10 | 10 | 10 | 10 | — | — | — |