Binary Classification on STL10
98.74Top-1 AccuracyRefCal
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| RefCalBackbone=ResNet-50, Baseline=MMCE2026.05 | 98.74 | 99.88 | 1.13 | 1.16 | 1 | 0.98 | |
| RefCalBackbone=ResNet-50, Baseline=MbLS2026.05 | 98.71 | 99.91 | 0.82 | 0.86 | 0.64 | 0.85 | |
| RefCalBackbone=ResNet-50, Baseline=NLL (CE)2026.05 | 98.69 | 99.8 | 1.13 | 1.11 | 1.02 | 1 | |
| RefCalBackbone=ResNet-50, Baseline=CE + TS2026.05 | 98.69 | 99.9 | 0.26 | 0.51 | 0.46 | 0.47 | |
| RefCalBackbone=ResNet-50, Baseline=CRL2026.05 | 98.69 | 99.88 | 1.13 | 1.12 | 1.02 | 1.01 | |
| RefCalBackbone=ResNet-50, Baseline=AdaFocal2026.05 | 98.69 | 98.68 | 1.31 | 0.67 | 0.64 | 0.67 | |
| RefCalBackbone=ResNet-50, Baseline=FL + MDCA2026.05 | 98.68 | 99.35 | 1.3 | 1.28 | 1.12 | 0.8 | |
| RefCalBackbone=ResNet-50, Baseline=LS2026.05 | 98.58 | 99.9 | 3.02 | 3.21 | 2.97 | 3.03 | |
| RefCalBackbone=ResNet-50, Baseline=MixUp2026.05 | 98.58 | 99.9 | 3.16 | 3.17 | 3.13 | 3.16 | |
| RefCalBackbone=ResNet-50, Baseline=AUCM2026.05 | 98.56 | 99.9 | 0.56 | 0.81 | 0.82 | 0.64 | |
| MbLSBackbone=ResNet-502026.05 | 96.98 | 99.42 | 1.09 | 1.18 | 0.85 | 1.15 | |
| CRLBackbone=ResNet-502026.05 | 96.88 | 99.47 | 1.46 | 1.45 | 1.4 | 1.5 | |
| NLL (CE)Backbone=ResNet-502026.05 | 96.71 | 99.37 | 1.21 | 1.64 | 1.47 | 1.2 | |
| CE + TSBackbone=ResNet-502026.05 | 96.71 | 99.4 | 0.41 | 1.21 | 1.16 | 0.66 | |
| LSBackbone=ResNet-502026.05 | 96.64 | 98.92 | 9.23 | 9.25 | 9.23 | 9.32 | |
| MixUpBackbone=ResNet-502026.05 | 96.58 | 99.36 | 6.01 | 6.3 | 6.26 | 6.04 | |
| MMCEBackbone=ResNet-502026.05 | 96.19 | 99.3 | 1.22 | 1.38 | 1.19 | 1.22 | |
| AdaFocalBackbone=ResNet-502026.05 | 96.16 | 99.11 | 1.1 | 1.1 | 1.03 | 1.06 | |
| AUCMBackbone=ResNet-502026.05 | 95.49 | 98.86 | 3.04 | 3.11 | 2.97 | 2.76 | |
| FL + MDCABackbone=ResNet-502026.05 | 95.49 | 98.23 | 0.93 | 1.97 | 1.98 | 0.98 |