Image Classification on CIFAR-10 (test) (Top-1 and Top-10 Accuracy)
88.5Top-1 AccFedAvg
Evaluation results
| Method | Links | ||
|---|---|---|---|
| FedAvgModel=ResNet-9, Participation (p)=0.1, Heterogeneity (gamma)=0.5, R (Total Rounds)=200, T (Local Steps)=5, eta_l (Learning Rate)=0.12026.01 | 88.5 | — | |
| SA-PEFModel=ResNet-9, Participation (p)=0.1, Heterogeneity (gamma)=0.5, R (Total Rounds)=200, T (Local Steps)=5, eta_l (Learning Rate)=0.12026.01 | 80.5 | 82.7 | |
| SAEFModel=ResNet-9, Participation (p)=0.1, Heterogeneity (gamma)=0.5, R (Total Rounds)=200, T (Local Steps)=5, eta_l (Learning Rate)=0.12026.01 | 74.2 | 75.5 | |
| FedAvgModel=ResNet-9, Participation (p)=0.5, Heterogeneity (gamma)=0.1, R (Total Rounds)=200, T (Local Steps)=5, eta_l (Learning Rate)=0.12026.01 | 69.5 | — | |
| CSERModel=ResNet-9, Participation (p)=0.1, Heterogeneity (gamma)=0.5, R (Total Rounds)=200, T (Local Steps)=5, eta_l (Learning Rate)=0.12026.01 | 68.6 | 80.5 | |
| EFModel=ResNet-9, Participation (p)=0.1, Heterogeneity (gamma)=0.5, R (Total Rounds)=200, T (Local Steps)=5, eta_l (Learning Rate)=0.12026.01 | 57 | 72.7 | |
| CSERModel=ResNet-9, Participation (p)=0.5, Heterogeneity (gamma)=0.1, R (Total Rounds)=200, T (Local Steps)=5, eta_l (Learning Rate)=0.12026.01 | 49.5 | 67.2 | |
| SA-PEFModel=ResNet-9, Participation (p)=0.5, Heterogeneity (gamma)=0.1, R (Total Rounds)=200, T (Local Steps)=5, eta_l (Learning Rate)=0.12026.01 | 47.5 | 68.5 | |
| EFModel=ResNet-9, Participation (p)=0.5, Heterogeneity (gamma)=0.1, R (Total Rounds)=200, T (Local Steps)=5, eta_l (Learning Rate)=0.12026.01 | 40.3 | 47.2 | |
| SAEFModel=ResNet-9, Participation (p)=0.5, Heterogeneity (gamma)=0.1, R (Total Rounds)=200, T (Local Steps)=5, eta_l (Learning Rate)=0.12026.01 | 39.5 | 48.5 |