fixed: the second update in Per-FedAvg(FO)

CharlieDinh / pFedMe

Personalized Federated Learning with Moreau Envelopes (pFedMe) using Pytorch (NeurIPS 2020)

290 stars 88 forks source link

fixed: the second update in Per-FedAvg(FO) #9

Closed sshpark closed 3 years ago

sshpark commented 3 years ago

In the second update step, the model parameters is the one before first update step. Here, I fix the problem by restore it before applying gradient descent to it.