FedProx: number of epochs per round for collaborator inner optimizaton
chinglamchoi opened this issue · comments
Ching Lam Choi commented
Hello, thanks for the great work!
I've looked into the FedProx implementation but couldn't find the hyperparameter for customising the number of epochs for the inner optimization (of the invidual collaborators) in each round. In the original FedProx implementation, a federated experiment might contain 200 rounds, 10 clients per round, where each client is optimized for 20 epochs. How do I replicate this with openfl?
Thanks a lot!