How To Get Started With Data Science Workbench

Where are the best resources for getting started with Data Science Workbench?

Thank you,

– Rich

Welcome to the community Rich.

Have you looked at Overview — Workbench documentation ?


Yes, I have looked at the documentation, and it shows me how to install the data science workbench, but I can’t seem to find much guidance about how to make effective use of it. For example, see the other question I posted in the forum.

– Rich

Yes… I saw the other question. I think I know the answer, but want to be certain before I reply. As for getting the most out of Workbench, there is a GTC session on this. I’ll try to find a link to the session and include it when answering your other question… likely tomorrow.

To do what you are asking is a little bit complicated but it is possible with an SSH tunnel. Using the tunnel map the jupyter notebook port, should be 49153, running on the DSW host to a local port say 8080. Assuming you have already used the NV Workbench launcher to start a Pytorch or Tensorflow container, but any container will work. In a terminal window on the NV Workbench machine run,
ssh -L 49153:localhost:8080 <REMOTE_USER>@<REMOTE_HOST>
This command will open an ssh terminal to the remote machine. On Windows and Mac you’ll need to configure openssh. On the machine specified in REMOTE_HOST above in a web browser just access the jupyter notebook http://localhost:8080/ URL. The key to this is the running SSH tunnel ssh -L .

I’ll try to get back to you with a link to the Workbench GTC session.

Here is a link to a November 2021 GTC session that discusses the use of NVIDIA Data Science Workbench. Three Ways NVIDIA Improves the Data Science Experience | NVIDIA On-Demand

Sorry… one more thing. Not that this is important, but there is a 90 second overview of Workbench on YouTube: Faster Data Science Development with NVIDIA Data Science Workbench - YouTube