Image Classification on Indoor 25.18% correlated noise (test)
70.12Top-1 AccuracyREGSL
Evaluation Results
| Method | Links | |
|---|---|---|
| REGSLBackbone=ResNet-18, Evaluation Protocol=Fine-tuning, Averaging across random seeds=32021.11 | 70.12 | |
| ELRBackbone=ResNet-18, Evaluation Protocol=Fine-tuning, Averaging across random seeds=32021.11 | 69.38 | |
| SCEBackbone=ResNet-18, Evaluation Protocol=Fine-tuning, Averaging across random seeds=32021.11 | 69 | |
| GCEBackbone=ResNet-18, Evaluation Protocol=Fine-tuning, Averaging across random seeds=32021.11 | 68.88 | |
| l2-PGMBackbone=ResNet-18, Evaluation Protocol=Fine-tuning, Averaging across random seeds=32021.11 | 68.86 | |
| APLBackbone=ResNet-18, Evaluation Protocol=Fine-tuning, Averaging across random seeds=32021.11 | 68.31 | |
| LSBackbone=ResNet-18, Evaluation Protocol=Fine-tuning, Averaging across random seeds=32021.11 | 68.06 | |
| DACBackbone=ResNet-18, Evaluation Protocol=Fine-tuning, Averaging across random seeds=32021.11 | 68.06 | |
| Fine-tuningBackbone=ResNet-18, Evaluation Protocol=Fine-tuning, Averaging across random seeds=32021.11 | 67.49 | |
| SATBackbone=ResNet-18, Evaluation Protocol=Fine-tuning, Averaging across random seeds=32021.11 | 67.06 |