FedPer准确率问题 - Githubissues

TsingZ0 / FedALA

AAAI 2023 accepted paper, FedALA: Adaptive Local Aggregation for Personalized Federated Learning

Apache License 2.0

113 stars 18 forks source link

Closed XZssLYZCXrjD1l closed 2 weeks ago

XZssLYZCXrjD1l commented 2 months ago

哥，请问为啥我的FedPer准确率比FedRep还要高5个百分点，在50个客户端，Cifar100数据集，dir=0.1里。客户端的joinrate不管多少都高

TsingZ0 commented 1 month ago

请问具体的实验设置是？