Image Classification on CIFAR-100 (Loss and 25-Epoch Accuracy)
1.62Train LossW-Asymmetric ResNet20 (BN)
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| W-Asymmetric ResNet20 (BN)Model=ResNet20, Normalization=Batch Normalization2024.05 | 1.62 | 1.5 | 7.6 | 58.4 | 49.29 | |
| ResNet20 (BN)Model=ResNet20, Normalization=Batch Normalization2024.05 | 1.68 | 1.57 | 7.8 | 56.83 | 46.8 | |
| W-Asymmetric ResNet20 (LN)Model=ResNet20, Normalization=Layer Normalization2024.05 | 1.91 | 1.82 | 8.6 | 51.2 | 39.03 | |
| ResNet20 (LN)Model=ResNet20, Normalization=Layer Normalization2024.05 | 1.97 | 1.88 | 9 | 50.02 | 37.24 |