Model Training on CIFAR-100 (non-IID)
2,484.79Training Delay (minutes)Proposed
Evaluation results
| Method | Links | |
|---|---|---|
| ProposedModel=ResNet182025.06 | 2,484.79 | |
| ProposedModel=GoogLeNet2025.06 | 2,660.29 | |
| ProposedModel=ResNet502025.06 | 2,819.09 | |
| ProposedModel=DenseNet1212025.06 | 3,232.6 | |
| RegressionModel=ResNet182025.06 | 3,318.7 | |
| OSSModel=ResNet182025.06 | 3,327.46 | |
| Device-onlyModel=ResNet182025.06 | 3,361.4 | |
| OSSModel=GoogLeNet2025.06 | 3,601.12 | |
| OSSModel=ResNet502025.06 | 3,842.83 | |
| RegressionModel=ResNet502025.06 | 4,038.32 | |
| Device-onlyModel=ResNet502025.06 | 4,057.17 | |
| Device-onlyModel=GoogLeNet2025.06 | 4,105.14 | |
| OSSModel=DenseNet1212025.06 | 4,300.22 | |
| RegressionModel=GoogLeNet2025.06 | 4,390.73 | |
| RegressionModel=DenseNet1212025.06 | 4,954.29 | |
| Device-onlyModel=DenseNet1212025.06 | 5,021.96 |