Table 2.

Application used for the evaluation of EdgeMap.

Applications	MNIST-MLP	MNIST-LeNet	Fashion-MNIST	Heart Class	CIFAR10-LeNet	CIFAR10-AlexNet	CIFAR10-VGG11	CIFAR10-ResNet
	$28\times 28\times 1$	$28\times 28\times 1$	$28\times 28\times 1$	$42\times 42\times 1$	$32\times 32\times 3$	$32\times 32\times 3$	$32\times 32\times 3$	$32\times 32\times 3$
Topology	$MLP$ 1	$CNN$ 2	$MLP$ 3	$CNN$ 4	$CNN$ 5	$CNN$ 6	$CNN$ 7	$CNN$ 8
Neuron number	1193	7403	1393	17,001	11,461	794,232	9,986,862	9,675,543
Synapses	97,900	377,990	444,600	773,112	804,614	39,117,304	47,737,200	48,384,204
Total spikes	358,000	1,555,986	10,846,940	2,209,232	7,978,094	574,266,873	796,453,842	5,534,290,865

¹ Feedforward(784-100-10). ² Conv((5,5),(1,1),6)-AvgPool(2,2)-Conv((5,5),(1,1),16)-AvgPool(2,2)-FC(500)-FC(10). ³ Feedforward(784-500-100-10). ⁴ Conv((5,5),(1,1),6)-AvgPool(2,2)-Conv((5,5),(1,1),16)-AvgPool(2,2)-FC(10). ⁵ Conv((5,5),(1,1),6)-AvgPool(2,2)-Conv((5,5),(1,1),16)-AvgPool(2,2)-FC(500)-FC(10). ⁶ Conv((11,11),(4,4),96)-Maxpool(2,2)-Conv((5,5),(1,1),256)-Maxpool(2,2)-Conv((3,3),(1,1),384)-Conv((3,3),(1,1),256)-Maxpool(2,2)-FC(4096-4096-10). ⁷ Conv((3,3),(1,1),64)-MaxPool(2,2)-Conv((3,3),(1,1),128)-MaxPool(2,2)-Conv((3,3),(1,1),256)-Conv((3,3),(1,1),256)-MaxPool(2,2)-Conv((3,3),(1,1),512)-Conv((3,3),(1,1),512)-MaxPool(2,2))-Flatten-FC(4096-4096-10). ⁸ Conv((3,3),(1,1),64)-MaxPool(2,2)-Conv((3,3),(1,1),128)-MaxPool(2,2)- Conv((3,3),(1,1),256)-Conv((3,3),(1,1),256)-MaxPool(2,2)-Conv((3,3),(1,1),512)-Conv((3,3),(1,1),512)-MaxPool(2,2)- FC(4096-4096-10).