Image Classification on CIFAR-100 (test) (Type I, II, III Noise Robustness)
79.27Accuracy (0% noise)GJS
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| GJSBackbone=ResNet-34, Number of runs=5, Hyperparameter optimization=Standard setup2021.05 | 79.27 | 68.49 | 69.21 | 69.04 | |
| GCEBackbone=ResNet-34, Number of runs=5, Hyperparameter optimization=Standard setup2021.05 | 77.65 | 63.51 | 65.62 | 65.85 | |
| CEBackbone=ResNet-34, Number of runs=5, Hyperparameter optimization=Standard setup2021.05 | 77.6 | 62.46 | 65.84 | 62.44 |