Hands-On Deep Learning Algorithms with Python

Computing primary and digit capsules

Now we will compute the primary capsules, which extract the basic features, and the digit capsules, which recognizes the digits.

Start the TensorFlow Graph:

graph = tf.Graph()with graph.as_default() as g:

Define the placeholders for input and output:

    x = tf.placeholder(tf.float32, [batch_size, 784])    y = tf.placeholder(tf.float32, [batch_size,10])    x_image = tf.reshape(x, [-1,28,28,1])

Perform the convolution operation and get the convolutional input:

    with tf.name_scope('convolutional_input'):        input_data = tf.contrib.layers.conv2d(inputs=x_image, num_outputs=256, kernel_size=9, padding='valid')

Compute the primary capsules that extract the basic features, such as edges. First, compute the capsules using ...