- TensorFlow operations form a computation graph.
- Visualizing the graph can help both in diagnosing issues with the computation itself, but also in understanding how certain operations in TensorFlow work and how are things put together.
- We’ll take a look at a few different ways of visualizing TensorFlow graphs, and most importantly, show how to do it in a very simple and time-efficient way.
- Now onto the specifics, we’ll take a look at the following visualization techniques:
First, let us create a simple TensorFlow graph.
- Regular operations such as creating a placeholder with tf.placeholder will create a node in the so called default graph.
Prerequisites: This article assumes you are familiar with the basics of Python, TensorFlow, and Jupyter notebooks. We won’t use any of the advanced TensorFlo…
Continue reading “Visualizing TensorFlow Graphs in Jupyter Notebooks”