Deep Learning for the Classification of Genomic Signals

Table 1

Description of the layers for the proposed CNN architecture.

Type of layer	Patch size	Output size	Number of Parameters
Input			0
Convolution			896
Convolution
Max pooling			0
Convolution
Convolution
Max pooling			0
Convolution
Convolution
Convolution
Max pooling			1
Inception
Inception
Inception
Inception
Average pooling			0
Fully connected
Fully connected
Fully connected
Softmax			258