Image Classification on EMNIST Practical heterogeneous (alpha = 0.1, K = 20) (test)
96.84AccuracyCAFedCL
Evaluation results
| Method | Links | ||
|---|---|---|---|
| CAFedCLBackbone=ResNet-18, Normalization=Group Normalization, Local Training Epochs=5, Optimizer=SGD, Number of runs=52026.03 | 96.84 | 1.54 | |
| FedProtoBackbone=ResNet-18, Normalization=Group Normalization, Local Training Epochs=5, Optimizer=SGD, Number of runs=52026.03 | 95.5 | 1.69 | |
| FedTGPBackbone=ResNet-18, Normalization=Group Normalization, Local Training Epochs=5, Optimizer=SGD, Number of runs=52026.03 | 94.47 | 4.54 | |
| MP-FedCLBackbone=ResNet-18, Normalization=Group Normalization, Local Training Epochs=5, Optimizer=SGD, Number of runs=52026.03 | 94.3 | 3.28 | |
| FedRCLBackbone=ResNet-18, Normalization=Group Normalization, Local Training Epochs=5, Optimizer=SGD, Number of runs=52026.03 | 93.57 | 3.33 | |
| MOONBackbone=ResNet-18, Normalization=Group Normalization, Local Training Epochs=5, Optimizer=SGD, Number of runs=52026.03 | 92.77 | 6.53 | |
| FedLCBackbone=ResNet-18, Normalization=Group Normalization, Local Training Epochs=5, Optimizer=SGD, Number of runs=52026.03 | 90.45 | 7.08 | |
| FedProxBackbone=ResNet-18, Normalization=Group Normalization, Local Training Epochs=5, Optimizer=SGD, Number of runs=52026.03 | 84.27 | 8.13 | |
| FedAvgBackbone=ResNet-18, Normalization=Group Normalization, Local Training Epochs=5, Optimizer=SGD, Number of runs=52026.03 | 82.95 | 11.59 | |
| FedProcBackbone=ResNet-18, Normalization=Group Normalization, Local Training Epochs=5, Optimizer=SGD, Number of runs=52026.03 | 65.66 | 3.57 |