Open Warrior0x1 opened 6 days ago
Sorry I am away from my workstation now. I will evaluate pFedSim later when I am back. Please stay tuned.
What about other datasets? Is pFedSim only performs bad on medmnistA?
FedAvg is not performing as well on CIFAR-10 as it used to. I haven't tested the others.
Why does running pFedSim and FedAvg using the medmnistA dataset in your latest repository result in severe fluctuations and non-convergence? This issue didn't occur in previous repositories. Could it be a problem with my dataset settings? My configuration is as follows: