Image Classification on CIFAR-100 (Accuracy, Top-5, NLL, ECE, Brier, AUROC)
67.4AccuracylCBOpt-adapt
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| lCBOpt-adapt2025.11 | 67.4 | 90.7 | 1.268 | 11 | 0.457 | 84.5 | |
| uCBOptϑ=5 ⋅ 10−52025.11 | 67.1 | 90.6 | 1.231 | 9.5 | 0.453 | 84.7 | |
| uCBOptϑ=→ 0+2025.11 | 66.7 | 90.5 | 1.247 | 9.4 | 0.457 | 84.5 | |
| uCBOpt-adapt2025.11 | 65.5 | 89.7 | 1.298 | 9.6 | 0.475 | 84.1 | |
| SGD2025.11 | 64.9 | 89.1 | 1.334 | 9.6 | 0.483 | 83.8 | |
| MC-D2025.11 | 64.9 | 89.4 | 1.298 | 7.4 | 0.476 | 83.8 | |
| IVON2025.11 | 60.7 | 87 | 1.458 | 7.1 | 0.523 | 83 | |
| IVON@mean2025.11 | 59.9 | 86.4 | 1.521 | 9.2 | 0.538 | 82.8 |