Image Classification on CIFAR-10 Schematic confusion Standard
58.65AccuracyREFINE
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| REFINEBackbone=CNNs2025.05 | 58.65 | 0.9034 | 0.5861 | 38.4 | |
| DistillBackbone=CNNs2025.05 | 57.8 | 0.9068 | 0.5772 | 35.92 | |
| NoTransBackbone=CNNs2025.05 | 56.53 | 0.9006 | 0.5639 | 35.76 | |
| LoRABackbone=CNNs2025.05 | 49.96 | 0.9039 | 0.4864 | 16.34 | |
| DANN-GateBackbone=CNNs2025.05 | 49.04 | 0.9028 | 0.4719 | 11.4 | |
| LinearProbeBackbone=CNNs2025.05 | 48.54 | 0.8987 | 0.4757 | 18.44 | |
| AdapterBackbone=CNNs2025.05 | 47.17 | 0.8998 | 0.4479 | 7.42 |