Skip to main content

TensorBoard

This guide explains how to use TensorBoard within Gradient Notebooks.

Introduction to TensorBoard

TensorBoard is a visualization toolkit from TensorFlow. It is useful for a variety of applications including visualizing metrics and histograms, displaying images, and more.

Example TensorBoard

To get started with TensorBoard, we'll need to run a command in the terminal and then access a special URL.

How to configure a TensorBoard in Gradient Notebooks

To run TensorBoard in a notebook, first make sure that the notebook has TensorBoard installed.

Next, open a notebook terminal and run the following command:

tensorboard --logdir . --bind_all

If needed, the --logdir path may be changed to a different value.

The TensorBoard is now available at the following URL:

https://tensorboard-NOTEBOOKID.CLUSTERID.paperspacegradient.com

NOTEBOOKID may be found on the notebook list view in the Gradient console.

The default CLUSTERID value is clg07azjl. Private clusters will have a CLUSTERID on the clusters page.