Creating the first fully connected layer

In this recipe, let's create the first fully connected layer.

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 flattened convolution layer; that is, layer_flat
• Num_inputs: This is the number of features created post flattening, num_features
• Num_outputs: This is the number of fully connected neurons output, fc_size
• Use_relu: This is the binary flag set to TRUE to incorporate non-linearity in the tensor

How to do it...

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

layer_fc1 = create_fc_layer(input=layer_flat,
num_inputs=num_features,
num_outputs=fc_size,
use_relu=TRUE)

How it works...

Here, we create a fully connected layer that returns a two-dimensional tensor. The first dimension (?) represents any number of (input) images and the second dimension represents the number of output neurons (here, 1,024).