Table 13.

32 × 32 network architecture

Operation	Kernel	Strides	Feature maps	BN?	Dropout	Nonlinearity
$G_X\left(z\right)-110\times 1\times 1\text{input}$
Linear	N/A	N/A	384	$\times$	0.0	ReLU
Transposed convolution	$5\times 5$	$2\times 2$	192	$\surd$	0.0	ReLU
Transposed convolution	$5\times 5$	$2\times 2$	96	$\surd$	0.0	ReLU
Transposed convolution	$5\times 5$	$2\times 2$	1	$\times$	0.0	Tanh
$D\left(x\right)-32\times 32\times 1\text{input}$
Convolution	$3\times 3$	$2\times 2$	64	$\times$	0.5	Leaky ReLU
Convolution	$3\times 3$	$2\times 2$	128	$\surd$	0.5	Leaky ReLU
Convolution	$3\times 3$	$2\times 2$	256	$\surd$	0.5	Leaky ReLU
Convolution	$3\times 3$	$2\times 2$	512	$\surd$	0.5	Leaky ReLU
Convolution	$3\times 3$	$2\times 2$	512	$\surd$	0.5	Leaky ReLU
Linear	N/A	N/A	1	$\times$	0.0	Soft-sigmoid
Generator optimizer	Adam (α = [0.0001, 0.0002, 0.0003], β1 = 0.5, β2 = 0.999)
Discriminator optimizer	Adam (α = [0.0001, 0.0002, 0.0003], β1 = 0.5, β2 = 0.999)
Batch size	100
Iterations	15,000
Leaky ReLU slope	0.2
Activation noise standard deviation	[0, 0.1, 0.2]
Weight, bias initialization	Isotropic gaussian (µ = 0, σ = 0.02), constant (0)