Launching TensorBoard from a notebook

[sciaba (Andrea Sciaba)]

Hi,
I tried to launch TensorBoard from a Python notebook by doing

%load_ext tensorboard
%tensorboard --logdir=./my_logs --port=6006

but I just get a big empty cell with the message “swan005.cern.ch took too long to respond”.

Is this normal?

Thanks,
Andrea

[etejedor (Enric Tejedor Saavedra)]

Hello,

Since tensorboard is listening on a port that is not open in your session, that unfortunately can’t work.

There was a discussion about this topic here:

Tensorboard access

Tensorflow offers a nice visualisation tool for the model parameters: TensorBoard  |  TensorFlow It can be launched directly from a notebook (or from the CLI), but then one needs to access it on another port. Is there any way to make this work with SWAN? Of course, the current alternative is just to launch the tensorboard locally using the synce’d cernbox folder.

It seems there is a JLab extension we could use that would act as a proxy for tensorboard, but this is not yet in production.