Image Classification on CIFAR-100 (test) (Accuracy, Loss, ECE, Agr, Var, ODD)
68.49AccuracyDropout (SE1)
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Dropout (SE1)type=Monte Carlo Dropout, drop_rate=0.2, ensemble_size=502022.12 | 68.49 | 1.167 | 0.124 | 77.2 | 22.2 | 73.5 | |
| HMC 3-chaintype=Hamiltonian Monte Carlo, chains=32022.12 | 67.39 | 1.205 | 0.131 | 100 | 0 | 72.5 | |
| Regulartype=Deep Ensemble2022.12 | 62.84 | 1.477 | 0.157 | 71.4 | 29.8 | 71.5 | |
| Multi SWAtype=Stochastic Weight Averaging2022.12 | 61.86 | 1.526 | 0.164 | 70.3 | 31.3 | 71.6 |