Image Classification on CIFAR-LT-100 Imbalance Factor 50 (test)
63.67Top-1 AccuracyGLMC + LE-SAM
Evaluation results
| Method | Links | ||
|---|---|---|---|
| GLMC + LE-SAM2026.05 | 63.67 | — | |
| GLMC + SAM2026.05 | 63.01 | — | |
| GLMC2026.05 | 61.1 | — | |
| ResComBackbone=ResNet-322022.03 | 58 | — | |
| Weight Balancingregularization=+ WD & Max2022.03 | 57.71 | — | |
| Weight Balancingregularization=+ tau-norm2022.03 | 57.65 | — | |
| DRO-LT2022.03 | 57.57 | — | |
| Weight Balancingregularization=+ WD2022.03 | 57.47 | — | |
| RBL2026.05 | 57.2 | — | |
| Weight Balancingregularization=+ L2norm2022.03 | 56.33 | — | |
| Weight Balancingregularization=+ Max2022.03 | 56.06 | — | |
| PaCo2022.03 | 56 | — | |
| PaCoBackbone=ResNet-322022.03 | 56 | — | |
| Balanced SoftmaxBackbone=ResNet-322022.03 | 54.2 | — | |
| Weight Balancingregularization=WD2022.03 | 52.71 | — | |
| tau-norm2022.03 | 52.53 | — | |
| MiSLASBackbone=ResNet-322022.03 | 52.3 | — | |
| ACEnumber of experts=42022.03 | 51.9 | — | |
| HCL2026.05 | 51.9 | — | |
| RIDEBackbone=ResNet-32, experts=32022.03 | 51.7 | — | |
| DiVE2022.03 | 51.13 | — | |
| DiVEImbalance Factor=502021.03 | 51.13 | — | |
| SSD2022.03 | 50.5 | — | |
| LADEBackbone=ResNet-322022.03 | 50.5 | — | |
| ETF-DR2026.05 | 50.4 | — | |
| De-confound2022.03 | 50.3 | — | |
| Causal NormBackbone=ResNet-322022.03 | 50.3 | — | |
| TDEImbalance Factor=502021.03 | 50.3 | — | |
| Meta-learningImbalance Factor=502021.03 | 50.08 | — | |
| ResLTBackbone=ResNet-322022.03 | 50 | — | |
| SEL2026.05 | 49.6 | — | |
| Hybrid-PSCBackbone=ResNet-322022.03 | 49 | — | |
| Focal-SAM2026.05 | 48.1 | — | |
| BSCEImbalance Factor=502021.03 | 47.6 | — | |
| TSC2026.05 | 47.4 | — | |
| LDAM+SSP2022.03 | 47.11 | — | |
| LDAM-DRW+SSPImbalance Factor=502021.03 | 47.11 | — | |
| BBN2026.05 | 47.1 | — | |
| LogitAjust2022.03 | 47.03 | — | |
| BBN2022.03 | 47.02 | — | |
| BBNImbalance Factor=502021.03 | 47.02 | — | |
| BBNBackbone=ResNet-322022.03 | 47 | — | |
| LDAM-DRW2022.03 | 46.62 | — | |
| LDAM-DRWImbalance Factor=502021.03 | 46.62 | — | |
| KCL2026.05 | 46.3 | — | |
| CE2026.05 | 46.1 | — | |
| KD2022.03 | 45.49 | — | |
| CE+CB2022.03 | 45.32 | — | |
| Focal+CB2022.03 | 45.17 | — | |
| Focal2022.03 | 44.32 | — | |
| Focal+Imbalance Factor=502021.03 | 44.32 | — | |
| Weight Balancingregularization=naive2022.03 | 43.99 | — | |
| CE2022.03 | 43.85 | — | |
| CEImbalance Factor=502021.03 | 42.41 | — | |
| BBNBackbone=ResNet-32, Imbalance factor=502021.03 | — | 52.98 | |
| Class-balanced cross-entropy lossBackbone=ResNet-32, Imbalance factor=502021.03 | — | 55.21 | |
| Class-balanced fine-tuningBackbone=ResNet-32, Imbalance factor=502021.03 | — | 53.78 | |
| Class-balanced focal lossBackbone=ResNet-32, Imbalance factor=502021.03 | — | 54.79 | |
| Cross-entropy trainingBackbone=ResNet-32, Imbalance factor=502021.03 | — | 55.98 | |
| Focal lossBackbone=ResNet-32, Imbalance factor=502021.03 | — | 55.68 | |
| L2RWBackbone=ResNet-32, Imbalance factor=502021.03 | — | 56.83 | |
| LDAM lossBackbone=ResNet-32, Imbalance factor=502021.03 | — | 53.84 | |
| LDAM-DRWBackbone=ResNet-32, Imbalance factor=502021.03 | — | 52.03 | |
| Meta-class-weight with cross-entropy lossBackbone=ResNet-32, Imbalance factor=502021.03 | — | 51.47 | |
| Meta-class-weight with focal lossBackbone=ResNet-32, Imbalance factor=502021.03 | — | 49.92 | |
| Meta-class-weight with LDAM lossBackbone=ResNet-32, Imbalance factor=502021.03 | — | 50.84 | |
| Meta-weight netBackbone=ResNet-32, Imbalance factor=502021.03 | — | 54.34 | |
| MetaSAug with cross-entropy lossBackbone=ResNet-32, Imbalance factor=502021.03 | — | 48.1 | |
| MetaSAug with focal lossBackbone=ResNet-32, Imbalance factor=502021.03 | — | 48.38 | |
| MetaSAug with LDAM lossBackbone=ResNet-32, Imbalance factor=502021.03 | — | 47.73 | |
| MixupBackbone=ResNet-32, Imbalance factor=502021.03 | — | 55.01 |