PrettyTensor

PrettyTensor allows the developer to wrap TensorFlow operations to quickly chain any number of layers to define neural networks. Coming up is simple example of Pretty Tensor's capabilities: we wrap a standard TensorFlow object, pretty, into a library-compatible object; then we feed it through three fully connected layers, and we finally output a softmax distribution:

pretty = tf.placeholder([None, 784], tf.float32)
softmax = (prettytensor.wrap(examples)
 .fully_connected(256, tf.nn.relu)
 .fully_connected(128, tf.sigmoid)
 .fully_connected(64, tf.tanh)
 .softmax(10))

The PrettyTensor installation is very simple. You can just use the pip installer:

sudo pip install prettytensor

Chaining layers

PrettyTensor has three modes of operation that share the ability to chain methods.

Normal mode

In normal mode, every time a method is called, a new PrettyTensor is created. This allows easy chaining, and you can still use any particular object multiple times. This makes it easy to branch your...