A Homomorphic Signcryption-Based Privacy Preserving Federated Learning Framework for IoTs

Table 2

Datasets and corresponding models.

Datasets	Model structure	Model parameters (#)	Model size (MB)
CINIC-10	ResNet-56	591322	18.05
CIFAR-100	ResNet-56	591322	18.05
Federated EMNIST	CNN	1206590	36.82
CIFAR-10	ResNet-18+ group normalization	11232612	342.79

We should note that though ResNet-18 has less layers than ResNet-56; because of the group normalization operation, it has more parameters than ResNet-56.