Personalized Federated Learning with Semisupervised Distillation

Table 1

Comparison of the number of parameters of pFedSD-hot and baselines transmitted between the server and clients in each communication round.

Methods	The number parameters each communication round ×
FedMD	100
DS-FL	100
MHAT	100
pFedSD-soft	100
pFedSD-hot	55