Image Classification on CIFAR-100 (test) (Flip Noise Robustness)
70.65Accuracy (0% Noise)GDW
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| GDWNoise Type=Flip noise, Number of runs=52021.10 | 70.65 | 65.41 | 52.44 | |
| MWNetNoise Type=Flip noise, Number of runs=52021.10 | 70.64 | 64.72 | 50.62 | |
| Fine-tuningNoise Type=Flip noise, Number of runs=52021.10 | 70.52 | 65.02 | 50.23 | |
| INSWNoise Type=Flip noise, Number of runs=52021.10 | 70.52 | 65.32 | 50.13 | |
| Soft-labelNoise Type=Flip noise, Number of runs=52021.10 | 70.5 | 65.2 | 50.97 | |
| Gen-labelNoise Type=Flip noise, Number of runs=52021.10 | 70.46 | 64.94 | 49.93 | |
| BaseModelNoise Type=Flip noise, Number of runs=52021.10 | 70.42 | 64.96 | 49.83 | |
| Co-teachingNoise Type=Flip noise, Number of runs=52021.10 | 68.33 | 62.96 | 42.54 | |
| GLCNoise Type=Flip noise, Number of runs=52021.10 | 65.05 | 64.11 | 63.11 | |
| L2RWNoise Type=Flip noise, Number of runs=52021.10 | 63.4 | 55.27 | 45.41 |