Federated Learning on CIFAR-100 (100 clients, Dirichlet α=0.3)
49.56Accuracy (500 Rounds)FedACG
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| FedACGParticipation rate=2%, Dirichlet parameter=0.3, Network cost per round=1x2022.01 | 49.56 | 56.89 | 358 | |
| FedDCParticipation rate=2%, Dirichlet parameter=0.3, Network cost per round=2x2022.01 | 41.5 | 51.37 | 670 | |
| FedADAMParticipation rate=2%, Dirichlet parameter=0.3, Network cost per round=1x2022.01 | 40.95 | 51.14 | 592 | |
| FedMLBParticipation rate=2%, Dirichlet parameter=0.3, Network cost per round=1x2022.01 | 40.09 | 53.34 | 565 | |
| MOONParticipation rate=2%, Dirichlet parameter=0.3, Network cost per round=1x2022.01 | 39.91 | 46.51 | 730 | |
| FedLCParticipation rate=2%, Dirichlet parameter=0.3, Network cost per round=1x2022.01 | 39.7 | 42.1 | 1,000 | |
| FedDecorrParticipation rate=2%, Dirichlet parameter=0.3, Network cost per round=1x2022.01 | 39.42 | 48.45 | 718 | |
| FedAvgMParticipation rate=2%, Dirichlet parameter=0.3, Network cost per round=1x2022.01 | 39.24 | 53.47 | 504 | |
| FedDynParticipation rate=2%, Dirichlet parameter=0.3, Network cost per round=1x2022.01 | 38.94 | 48.88 | 716 | |
| FedAvgParticipation rate=2%, Dirichlet parameter=0.3, Network cost per round=1x2022.01 | 38.19 | 44.62 | 924 | |
| FedNTDParticipation rate=2%, Dirichlet parameter=0.3, Network cost per round=1x2022.01 | 38.13 | 48.03 | 708 | |
| FedProxParticipation rate=2%, Dirichlet parameter=0.3, Network cost per round=1x2022.01 | 36.69 | 45.16 | 921 | |
| FedCMParticipation rate=2%, Dirichlet parameter=0.3, Network cost per round=1.5x2022.01 | 14.52 | 23.06 | 1,000 |