Genetic CFL: Hyperparameter Optimization in Clustered Federated Learning

Table 4

Performance of FL against genetic CFL on CIFAR-10 dataset for various hyperparameters.

Client ratio	Rounds	FL		Genetic CFL
		Accuracy	Loss	Accuracy	Loss
0.10	3	0.6818	0.9540	0.6514	1.0097
	6	0.6891	1.3746	0.6639	1.3447
	10	0.6862	1.6617	0.6599	1.6449
0.15	3	0.6973	0.9612	0.7098	0.8814
	6	0.6988	1.3675	0.7199	1.693
	10	0.6952	1.6225	0.7129	1.3806
0.30	3	0.7578	0.8708	0.7688	0.7818
	6	0.7634	1.0891	0.7646	0.981
	10	0.7613	1.2964	0.7623	1.2961