Research on Fault Diagnosis Model of Generative Adss Based on Improved Semisupervised Diagnosis Algorithm

Table 2

Network structure parameters of CNN fault diagnosis model.

Layer (operation)	Input format	Output format
Convolution layer 1	[None,8,8,1]	[None,8,8,32]
Batch normalization	[None,8,8,32]	[None,8,8,32]
Nonlinear activation	[None,8,8,32]	[None,8,8,32]
Pool layer	[None,8,8,32]	[None,4,4,32]
Convolution layer 2	[None,4,4,32]	[None,4,4,64]
Batch normalization	[None,4,4,64]	[None,4,4,64]
Nonlinear activation	[None,4,4,64]	[None,4,4,64]
Pool layer 2	[None,4,4,64]	[None,2,2,64]
Fully connected layer	[None,256]	[None,6]
Softmax	[None,6]	[None,6]