Noisy-label detection on CIFAR-10
98.74Accuracy3D multi-metric
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| 3D multi-metricBackbone=Vit, Label error=10%2026.06 | 98.74 | 49.1 | 89.64 | |
| 3D multi-metricBackbone=Resnet50, Label error=40%2026.06 | 96.32 | 95.84 | 94.92 | |
| 2D geometric metricBackbone=Resnet50, Label error=40%2026.06 | 96.3 | 95.82 | 94.91 | |
| 3D multi-metricBackbone=Resnet50, Label error=20%2026.06 | 96.12 | 87.51 | 94.03 | |
| 2D geometric metricBackbone=Resnet50, Label error=20%2026.06 | 95.65 | 85.33 | 94.51 | |
| 2D geometric metricBackbone=Resnet50, Label error=10%2026.06 | 95.38 | 69.4 | 96.24 | |
| 2D geometric metricBackbone=Resnet50, Label error=5%2026.06 | 94.18 | 46.1 | 96.84 | |
| 3D multi-metricBackbone=Resnet50, Label error=10%2026.06 | 93.44 | 60.62 | 98.1 | |
| 2D geometric metricBackbone=Vit, Label error=10%2026.06 | 92.73 | 58.37 | 96.39 | |
| 2D geometric metricBackbone=Vit, Label error=20%2026.06 | 91.93 | 71.74 | 98.4 | |
| 3D multi-metricBackbone=Vit, Label error=20%2026.06 | 91.12 | 69.57 | 98.86 | |
| 3D multi-metricBackbone=Resnet50, Label error=5%2026.06 | 90.53 | 34.48 | 99.4 | |
| 2D geometric metricBackbone=Vit, Label error=5%2026.06 | 90.44 | 34.02 | 97.12 | |
| 3D multi-metricBackbone=Vit, Label error=40%2026.06 | 89.66 | 84.62 | 90.61 | |
| 2D geometric metricBackbone=Vit, Label error=40%2026.06 | 89.39 | 83.85 | 91 | |
| 3D multi-metricBackbone=Vit, Label error=5%2026.06 | 88 | 29.22 | 98.4 |