Intel® Distribution for Python*
Engage in discussions with community peers related to Python* applications and core computational packages.

2019.0 Docker image doesn't contain Jupyter

HimanshuG
Beginner
931 Views

Hello all,

The 2019.0 docker image of Intel Distribution for Python doesn't seem to contain Jupyter.  According to the announcement page for Docker Images for Intel® Distribution for Python, it should, as it provides an example on how to run it.

There is already one comment on the announcement page that it doesn't work, and I can confirm that. Is the document out of sync, or is it an oversight in the docker image?

Thanks!

 

 

0 Kudos
1 Solution
Todd_T_Intel
Employee
931 Views

Hello,

Jupyter was removed in the 2019.3 release because we have not provided any performance enhancements to Jupyter itself. You can install the Jupyter suite from the Anaconda defaults channel.

The documentation looks to be out-of-date as to which version is "latest". Did you use the "latest" tag, or specifically the 2019.0.0 tag?

Best, Todd

View solution in original post

0 Kudos
3 Replies
Todd_T_Intel
Employee
932 Views

Hello,

Jupyter was removed in the 2019.3 release because we have not provided any performance enhancements to Jupyter itself. You can install the Jupyter suite from the Anaconda defaults channel.

The documentation looks to be out-of-date as to which version is "latest". Did you use the "latest" tag, or specifically the 2019.0.0 tag?

Best, Todd

0 Kudos
HimanshuG
Beginner
931 Views

Hi @Todd T.

My original post had a mistake - it was the 2019.3 image, but as you said the documentation on the DockerHub page said 2019.0. Looks like you've updated it, and the 2019.4 image has also been pushed, so I'll pull latest now.  One more place that needs an update is the Docker image announcement link (in my original post) which shows an example of running Jupyter on the Docker image.

Too bad Jupyter has been removed, although I see the pattern Intel is using the decide what goes into the image.  So I guess I should probably be creating my own image using the Intel image as the base and adding in packages I need.  Are package removals between releases listed?

Thanks.

0 Kudos
Todd_T_Intel
Employee
931 Views

Hello,

Yes, creating your own image with Jupyter and any other packages you require with our image as the Base is a good model.

Jupyter is a fast-moving suite, so it is better for everyone to keep up to date by getting packages from the main distribution channels.

We document in the Release Notes when we add and remove packages. All our documentation can be found on our product page here: https://software.intel.com/en-us/distribution-for-python/documentation/view-all .

Best, Todd

0 Kudos
Reply