Difference between revisions of "Tensorboard"

From UFRC
Jump to navigation Jump to search
(Created page with "TensorBoard is used for visually monitoring statistics while a neural network trains. One recommendation for using [https://github.com/tensorflow/tensorboard/blob/master/docs...")
 
 
(6 intermediate revisions by 3 users not shown)
Line 1: Line 1:
 +
Back to [[TensorFlow]] page
 +
 
TensorBoard is used for visually monitoring statistics while a neural network trains.
 
TensorBoard is used for visually monitoring statistics while a neural network trains.
  
One recommendation for using [https://github.com/tensorflow/tensorboard/blob/master/docs/get_started.ipynb | TensorBoard] on HiPerGator is to run TensorBoard monitoring as a separate workload. Many neural network software frameworks are able to stream training data to a TensorBoard log file. Here are the steps to monitor the log separately from your training workload:
+
Our recommendation for using [https://github.com/tensorflow/tensorboard/blob/master/docs/get_started.ipynb TensorBoard] on HiPerGator is to run TensorBoard monitoring as a separate workload within a HiPerGator Desktop session. Many neural network software frameworks are able to stream training data to a TensorBoard log file.  
 +
 
 +
=== Here are the steps to monitor the log separately from your training workload===
 +
 
 +
* Request a '''HiPerGator Desktop''' via the '''Interactive Apps Menu''' of [https://ood.rc.ufl.edu OnDemand]
 +
** For resources, something like 1 CPU and 2 GB RAM should be good.
 +
* Connect to the session when it starts and open a "Terminal Emulator" from the desktop Applications menu in the top left.
 +
* Start a TensorBoard instance in the resulting terminal/shell (replace "./pathtolog" with the path to where your TensorFlow process is storing its logs):
 +
  $ module load tensorflow ubuntu
 +
  $ tensorboard --logdir ./pathtolog &
 +
 
 +
* After a few seconds, TensorBoard will display a message similar to this (There may be some errors displayed, but TensorBoard should still work):
 +
...
 +
TensorBoard 2.4.0 at '''<nowiki>http://localhost:6006/</nowiki>''' (Press CTRL-C to quit)
 +
 
 +
* Copy the  printed URL ("<nowiki>http://localhost:6006/</nowiki>" in the example above)
 +
 
 +
* Press <ENTER> to get back the '$' prompt
  
* Request a 1 CPU 2 GB RAM HiPerGator Desktop via ood.rc.ufl.edu
+
* Start a browser:
* Open a terminal
+
$ firefox (or chrome)
** Enter command: ml tensorflow
 
** Enter command: tensorboard --logdir ./pathtolog
 
** See printed information including a localhost URL with port number
 
  
* Open a second terminal
+
* Paste the TensorBoard URL into the browser's location bar to view the TensorBoard UI.
** Enter command: ml ubuntu
 
** Enter command: firefox (or chrome)
 
** Paste URL into browser to view TensorBoard log
 

Latest revision as of 21:03, 12 December 2023

Back to TensorFlow page

TensorBoard is used for visually monitoring statistics while a neural network trains.

Our recommendation for using TensorBoard on HiPerGator is to run TensorBoard monitoring as a separate workload within a HiPerGator Desktop session. Many neural network software frameworks are able to stream training data to a TensorBoard log file.

Here are the steps to monitor the log separately from your training workload

  • Request a HiPerGator Desktop via the Interactive Apps Menu of OnDemand
    • For resources, something like 1 CPU and 2 GB RAM should be good.
  • Connect to the session when it starts and open a "Terminal Emulator" from the desktop Applications menu in the top left.
  • Start a TensorBoard instance in the resulting terminal/shell (replace "./pathtolog" with the path to where your TensorFlow process is storing its logs):
  $ module load tensorflow ubuntu
  $ tensorboard --logdir ./pathtolog &
  • After a few seconds, TensorBoard will display a message similar to this (There may be some errors displayed, but TensorBoard should still work):
...
TensorBoard 2.4.0 at http://localhost:6006/ (Press CTRL-C to quit)
  • Copy the printed URL ("http://localhost:6006/" in the example above)
  • Press <ENTER> to get back the '$' prompt
  • Start a browser:
$ firefox (or chrome)
  • Paste the TensorBoard URL into the browser's location bar to view the TensorBoard UI.