Tensor Builder

TensorBuilder is light-weight extensible library that enables you to easily create complex deep neural networks through a functional fluent immutable API based on the Builder Pattern. Tensor Builder also comes with a DSL based on applicatives and function composition that enables you to express more clearly the structure of your network, make changes faster, and reuse code.

Goals

Be a light-wrapper around Tensor-based libraries
Enable users to easily create complex branched topologies while maintaining a fluent API (see Builder.branch)
Let users be expressive and productive through a DSL

Installation

Tensor Builder assumes you have a working tensorflow installation. We don't include it in the requirements.txt since the installation of tensorflow varies depending on your setup.

From pypi

pip install tensorbuilder==0.0.18

From github

For the latest development version

pip install git+https://github.com/cgarciae/tensorbuilder.git@develop

Getting Started

Create neural network with a [5, 10, 3] architecture with a softmax output layer and a tanh hidden layer through a Builder and then get back its tensor:

import tensorflow as tf
from tensorbuilder import tb

x = tf.placeholder(tf.float32, shape=[None, 5])
keep_prob = tf.placeholder(tf.float32)

h = (
  tb
  .build(x)
  .tanh_layer(10) # tanh(x * w + b)
  .dropout(keep_prob) # dropout(x, keep_prob)
  .softmax_layer(3) # softmax(x * w + b)
  .tensor()
)

Features

Branching: Enable to easily express complex complex topologies with a fluent API. See Branching.
Scoping: Enable you to express scopes for your tensor graph using methods such as tf.device and tf.variable_scope with the same fluent API. Scoping.
DSL: Use an abbreviated notation with a functional style to make the creation of networks faster, structural changes easier, and reuse code. See DSL.
Patches: Add functions from other Tensor-based libraries as methods of the Builder class. TensorBuilder gives you a curated patch plus some specific patches from TensorFlow and TFLearn, but you can build you own to make TensorBuilder what you want it to be. See Patches.

Documentation

The Guide

Check out The Guide to learn to code in TensorBuilder.

Full Example

Next is an example with all the features of TensorBuilder including the DSL, branching and scoping. It creates a branched computation where each branch is executed on a different device. All branches are then reduced to a single layer, but the computation is the branched again to obtain both the activation function and the trainer.

import tensorflow as tf
from tensorbuilder import tb

x = placeholder(tf.float32, shape=[None, 10])
y = placeholder(tf.float32, shape=[None, 5])

[activation, trainer] = tb.pipe(
    x,
    [
        { tf.device("/gpu:0"):
            tb.relu_layer(20)
        }
    ,
        { tf.device("/gpu:1"):
            tb.sigmoid_layer(20)
        }
    ,
        { tf.device("/cpu:0"):
            tb.tanh_layer(20)
        }
    ],
    tb.linear_layer(5),
    [
        tb.softmax() # activation
    ,
        tb
        .softmax_cross_entropy_with_logits(y) # loss
        .map(tf.train.AdamOptimizer(0.01).minimize) # trainer
    ],
    tb.tensors()
)

Introduction