Skip to main content
. 2022 Feb 22;2022:2213273. doi: 10.1155/2022/2213273

Table 9.

Accuracy and compression rate versus batch size for LeNet-5 architecture.

Batch size Dataset Alive CR Accuracy
50 MNIST 10322 1.30 97.88
CIFAR-10 22705 11.41 32.56

100 MNIST 10344 1.30 98.53
CIFAR-10 7646 33.87 16.26

150 MNIST 10274 1.30 88.67
CIFAR-10 8273 31.30 17.33

200 MNIST 10285 1.30 89.84
CIFAR-10 6989 37.06 21.15

250 MNIST 10300 1.30 98.74
CIFAR-10 6203 41.75 25.00