Skip to main content

Table 4 Neural network architecture

From: A novel protein descriptor for the prediction of drug binding sites

layer number structure of each layer kernel size layer input layer output
1 Conv3D,ReLU 2,(8, 8, 8) 4,(16,16,16) 2,(16,16,16)
2 Conv3D,ReLU 4,(8, 8, 8) 2,(16,16,16) 4,(16,16,16)
3 MaxPooling3D (2,2,2) 4,(16,16,16) 4,(8,8,8)
4 Dropout(0.25) NA 4,(8,8,8) 4,(8,8,8)
5 Conv3D,ReLU 8,(4, 4, 4) 4,(8,8,8) 8,(8,8,8)
6 Conv3D,ReLU 16,(4, 4, 4) 8,(8,8,8) 16,(8,8,8)
7 MaxPooling3D (2,2,2) 16,(8,8,8) 16,(4,4,4)
8 Conv3D,ReLU 32,(2, 2, 2) 16,(4,4,4) 32,(4,4,4)
9 Conv3D,ReLU 64,(2, 2, 2) 32,(4,4,4) 64,(4,4,4)
10 Dropout(0.25) NA 32,(4,4,4) 64,(4,4,4)
11 Flatten NA 32,(4,4,4) 64,(4,4,4)
12 Dense(128),ReLU NA 4096 128
13 Dropout(0.5) NA 128 128
14 Dense(1),sigmoid NA 128 1