Image Classification on CIFAR-100 (test) (Top-1 and Top-10 Accuracy)
62.5Top-1 AccuracyFedAvg
Evaluation results
| Method | Links | ||
|---|---|---|---|
| FedAvgModel=ResNet-18, Participation (p)=0.1, Heterogeneity (gamma)=0.5, R (Total Rounds)=200, T (Local Steps)=5, eta_l (Learning Rate)=0.12026.01 | 62.5 | — | |
| FedAvgModel=ResNet-18, Participation (p)=0.5, Heterogeneity (gamma)=0.1, R (Total Rounds)=200, T (Local Steps)=5, eta_l (Learning Rate)=0.12026.01 | 61.9 | — | |
| SA-PEFModel=ResNet-18, Participation (p)=0.1, Heterogeneity (gamma)=0.5, R (Total Rounds)=200, T (Local Steps)=5, eta_l (Learning Rate)=0.12026.01 | 49.6 | 54.5 | |
| SA-PEFModel=ResNet-18, Participation (p)=0.5, Heterogeneity (gamma)=0.1, R (Total Rounds)=200, T (Local Steps)=5, eta_l (Learning Rate)=0.12026.01 | 48.5 | 60.6 | |
| CSERModel=ResNet-18, Participation (p)=0.1, Heterogeneity (gamma)=0.5, R (Total Rounds)=200, T (Local Steps)=5, eta_l (Learning Rate)=0.12026.01 | 46.2 | 51.5 | |
| SAEFModel=ResNet-18, Participation (p)=0.1, Heterogeneity (gamma)=0.5, R (Total Rounds)=200, T (Local Steps)=5, eta_l (Learning Rate)=0.12026.01 | 46.1 | 50.5 | |
| SAEFModel=ResNet-18, Participation (p)=0.5, Heterogeneity (gamma)=0.1, R (Total Rounds)=200, T (Local Steps)=5, eta_l (Learning Rate)=0.12026.01 | 44.5 | 57.5 | |
| CSERModel=ResNet-18, Participation (p)=0.5, Heterogeneity (gamma)=0.1, R (Total Rounds)=200, T (Local Steps)=5, eta_l (Learning Rate)=0.12026.01 | 42.5 | 60.7 | |
| EFModel=ResNet-18, Participation (p)=0.5, Heterogeneity (gamma)=0.1, R (Total Rounds)=200, T (Local Steps)=5, eta_l (Learning Rate)=0.12026.01 | 41.5 | 57.5 | |
| EFModel=ResNet-18, Participation (p)=0.1, Heterogeneity (gamma)=0.5, R (Total Rounds)=200, T (Local Steps)=5, eta_l (Learning Rate)=0.12026.01 | 40.4 | 49.5 |