A CNN is composed of convolution layers (defined by tf.nn.conv2d), a non-linearity layer (tf.nn.relu), a max pool (tf.nn.max_pool), and fully connected layers (tf.matmul). The following image shows typical CNN layers and their corresponding implementations in TensorFlow:
CNN in TensorFlow
Mapping CNN layers to TensorFlow functions
Get Neural Network Programming with TensorFlow now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.