tfgraphviz

.. image:: https://img.shields.io/github/license/akimach/tfgraphviz.svg
:alt: GitHub license
:target: https://github.com/akimach/tfgraphviz/blob/master/LICENSE

.. image:: https://badge.fury.io/py/tfgraphviz.svg
:target: https://badge.fury.io/py/tfgraphviz

tfgraphviz is a module to visualize a TensorFlow's data flow graph like TensorBoard using Graphviz. tfgraphviz enables to provide a visualization of tensorflow graph on Jupyter Notebook without TensorBoard.

Links

• GitHub: http://github.com/akimach/tfgraphviz

• PyPI: https://pypi.python.org/pypi/tfgraphviz

• Jupyter Notebook:

  .. image:: https://mybinder.org/badge.svg
  :target: https://mybinder.org/v2/gh/akimach/tfgraphviz/master?filepath=examples%2Fjupyter_sample.ipynb

  .. image:: https://img.shields.io/badge/view%20on-nbviewer-brightgreen.svg
  :target: https://nbviewer.jupyter.org/github/akimach/tfgraphviz/blob/master/examples/jupyter_sample.ipynb

Installation

Use pip to install:

.. code-block:: bash

$ pip install graphviz
$ pip install tfgraphviz

The only dependency is Graphviz.

macOS:

.. code-block:: bash

$ brew install graphviz

Ubuntu:

.. code-block:: bash

$ apt-get install graphviz

Quickstart

.. code-block:: python

import tensorflow as tf
import tfgraphviz as tfg

g = tf.Graph()
with g.as_default():
    a = tf.constant(1, name="a")
    b = tf.constant(2, name="b")
    c = tf.add(a, b, name="add")
tfg.board(g)

.. image:: https://raw.githubusercontent.com/akimach/tfgraphviz/master/img/graph.jpg
:align: center

License

This package is distributed under the MIT license <https://raw.githubusercontent.com/akimach/tfgraphviz/master/LICENSE>_.

Author

Akimasa KIMURA <https://github.com/akimach>_