Image Classification on CIFAR-10 Symmetric noise 0.8 manually corrupted (test)
52.27AccuracyINCV
Evaluation results
| Method | Links | |
|---|---|---|
| INCVBackbone=ResNet-32, Number of runs=52019.05 | 52.27 | |
| DecouplingBackbone=ResNet-32, Number of runs=52019.05 | 36.9 | |
| F-correctionBackbone=ResNet-32, Number of runs=52019.05 | 34.76 | |
| MentorNetBackbone=ResNet-32, Number of runs=52019.05 | 28.89 | |
| Co-teachingBackbone=ResNet-32, Number of runs=52019.05 | 16.21 | |
| D2LBackbone=ResNet-32, Number of runs=52019.05 | 10.02 |