Zhong
Zhong
train_args: federated_optimizer: "FedAvg" client_id_list: "[1,2,3,4,5]" client_num_in_total: 5 client_num_per_round: 2 comm_round: 6 epochs: 3 batch_size: 10 client_optimizer: sgd learning_rate: 0.03 weight_decay: 0.001 in my yaml file, the client_num_in_total was 5, however,...
After checking the source code, I found that fedml does not provide an implementation interface for data segmentation and loading. The user must prepare the processed data by himself.