Table 4.

Hyperparameters for all the models

Architecture	Learning rate	Decay	Momentum	Dropout	Batch size	Epochs	Hidden unit
W-Net	0.0001			0.6	256	500
W-Net with SVM	0.0001			0.6	256	500
AlexNet	0.001	0.0005	0.9	0.5	128	90
VGGNet	0.000001			0.5	1	300
ResNet50	0.001	0.0001	0.9		32	50
ResNet18	0.001	0.0001	0.9		32	50
RNN	0.01				64		32