Image Classification on RxRx1 ID WILDS (val)
35.9Top-1 AccuracyBaseline
Evaluation results
| Method | Links | |
|---|---|---|
| BaselineBackbone=ResNet502021.04 | 35.9 | |
| BN adaptBackbone=ResNet502021.04 | 35 | |
| ENTBackbone=ResNet502021.04 | 34.8 | |
| RPLBackbone=ResNet502021.04 | 34.8 | |
| IID Representation LearningModel=ResNet-50, TTA=False2026.03 | 23.9 | |
| ERM (CutMix)Model=ResNet-50, TTA=False2026.03 | 23.6 | |
| ABRAModel=ResNet-50, TTA=False2026.03 | 22.6 | |
| ARM-BNModel=ResNet-50, TTA=False2026.03 | 20.9 | |
| LISAModel=ResNet-50, TTA=False2026.03 | 20.1 | |
| ERM (Grid Search)Model=ResNet-50, TTA=False2026.03 | 19.4 | |
| IRMXModel=ResNet-50, TTA=False2026.03 | 18.9 | |
| IRMX (PAIR Opt)Model=ResNet-50, TTA=False2026.03 | 18.8 | |
| CORALModel=ResNet-50, TTA=False2026.03 | 18.5 | |
| Group DROModel=ResNet-50, TTA=False2026.03 | 15.2 | |
| Test-time BN AdaptationModel=ResNet-50, TTA=False2026.03 | 12.3 | |
| IRMModel=ResNet-50, TTA=False2026.03 | 5.6 |