Creating the second fully connected layer with dropout

In this recipe, let's create the second fully connected layer along with dropout.

Getting ready

The following are the inputs to the function defined in the recipe Using functions to flatten the densely connected layer, create_fc_layer:

• Input: This is the output of the first fully connected layer; that is, layer_fc1
• Num_inputs: This is the number of features in the output of the first fully connected layer, fc_size
• Num_outputs: This is the number of the fully connected neurons output (equal to the number of labels, num_classes )
• Use_relu: This is the binary flag set to FALSE

How to do it...

1. Run the create_fc_layer function with the preceding input parameters:

layer_fc2 = create_fc_layer(input=layer_fc1_drop,
num_inputs=fc_size,
num_outputs=num_classes,
use_relu=FALSE)

2. Use TensorFlow's dropout function to handle the scaling and masking of neuron outputs:

layer_fc2_drop <- tf$nn$dropout(layer_fc2, keep_prob)

How it works...

In step 1, we create a...