Table 3.

Architecture and training parameters of the 2D convolutional neural network.

Layers and Training Parameters	Values
Input layer	8 × 50
1st 2D convolutional layer	16 filters, filter size 3 × 3, stride 1 × 1
2nd 2D convolutional layer	32 filters, filter size 3 × 3, stride 1 × 1
3rd 2D convolutional layer	64 filters, filter size 3 × 3, stride 1 × 1
All 2D max pooling layer	pooling size 2 × 2, stride 2 × 2
1st fully connected layer	256 units
2nd fully connected layer	27 units
dropout layer	50%
Initial learn rate	0.0016
Learning rate drop factor	0.5
Learn rate drop period	4
Max epochs	20