Classification on CIFAR-10 (Calibration Metrics)
0.0526N-ERRepVGG
Evaluation results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| RepVGGScore=raw2026.05 | 0.0526 | 0.0318 | 92.17 | 0.876 | 1.6854 | 0.0921 | 0.0889 | 0.61 | 0.1859 | 0.0936 | 0.053 | |
| RepVGGScore=cal2026.05 | 0.0528 | 0.0084 | 91.46 | 0.638 | 1.4953 | 0.0736 | 0.0824 | 0.69 | 0.139 | 0.0887 | 0.0532 | |
| VGG19Score=raw2026.05 | 0.0677 | 0.0504 | 92.09 | 1.234 | 1.8889 | 0.1528 | 0.1237 | 0.75 | 0.3118 | 0.1268 | 0.0682 | |
| VGG19Score=cal2026.05 | 0.0683 | 0.0128 | 91.84 | 0.6892 | 1.5651 | 0.1033 | 0.1103 | 0.81 | 0.1804 | 0.1165 | 0.0689 | |
| ResNet20Score=raw2026.05 | 0.0822 | 0.0382 | 92.16 | 0.7942 | 1.5977 | 0.1223 | 0.1319 | 0.92 | 0.2368 | 0.1407 | 0.0829 | |
| ResNet20Score=cal2026.05 | 0.0839 | 0.007 | 92.33 | 0.6123 | 1.436 | 0.1011 | 0.1243 | 0.93 | 0.19 | 0.1364 | 0.0845 |