Image Classification on CIFAR-10 (test) (Accuracy and Robustness)
91.1AccuracyMaCS
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MaCSBackbone=ResNet-502026.03 | 91.1 | 43.48 | |
| MixupBackbone=ResNet-502026.03 | 89.63 | 42.68 | |
| Label SmoothingBackbone=ResNet-502026.03 | 88.85 | 41.87 | |
| Baseline (CE)Backbone=ResNet-502026.03 | 87.63 | 39.92 | |
| Focal LossBackbone=ResNet-502026.03 | 87.35 | 40.94 |