301

How can I remove a kernel from Jupyter Notebook?

I have an R kernel on my Jupyter Notebook. Recently, the kernel always dies right after I open a new notebook.

8 Answers 8

540

Run jupyter kernelspec list to get the paths of all your kernels.
Then simply uninstall your unwanted-kernel

jupyter kernelspec uninstall unwanted-kernel

Old answer
Delete the folder corresponding to the kernel you want to remove.

The docs has a list of the common paths for kernels to be stored in: http://jupyter-client.readthedocs.io/en/latest/kernels.html#kernelspecs

Sign up to request clarification or add additional context in comments.

4 Comments

Maybe obvious comment: You have to activate the virtualenv in which you installed jupyter before you can run the command.
jupyter kernelspec remove <kernel_name> now exists see my answer below stackoverflow.com/a/52606602/4413446
Yes, if the kernel is deleted manually then that is effectively the same as uninstalling it
What if this leaves some traces? When I install a kernel with the same name after this, the kernel does not work properly.
90

jupyter kernelspec remove now exists, see #7934. So you can just.

# List all kernels and grap the name of the kernel you want to remove
jupyter kernelspec list
# Remove it
jupyter kernelspec remove <kernel_name>

That's it.

2 Comments

Yeah, uninstall is simply an alias of remove these days...
This is a more complete answer, the first command allows you to check all the venvs linked to the jupyter notebooks.
78

You can delete it in the terminal via:

jupyter kernelspec uninstall yourKernel

where yourKernel is the name of the kernel you want to delete.

3 Comments

sudo may not even be necessary. This is a much better answer than the accepted one.
I did it today without sudo. And definitely this is the cleanest way. This should be the accepted answer.
The accepted answer was updated at some point with my content, so that the user can keep collecting upvotes... :D
30

There are two ways which is what I found. Either go to the directory where kernels are residing and delete from there. Secondly, using this command below.

List all kernels and grab the name of the kernel you want to remove

 jupyter kernelspec list

to get the paths of all your kernels.

Then simply uninstall your unwanted-kernel

jupyter kernelspec remove kernel_name

Comments

15

Just for completeness, you can get a list of kernels with jupyter kernelspec list, but I ran into a case where one of the kernels did not show up in this list. You can find all kernel names by opening a Jupyter notebook and selecting Kernel -> Change kernel. If you do not see everything in this list when you run jupyter kernelspec list, try looking in common Jupyter folders:

ls ~/.local/share/jupyter/kernels  # usually where local kernels go
ls /usr/local/share/jupyter/kernels  # usually where system-wide kernels go
ls /usr/share/jupyter/kernels  # also where system-wide kernels can go

Also, you can delete a kernel with jupyter kernelspec remove or jupyter kernelspec uninstall. The latter is an alias for remove. From the in-line help text for the command:

uninstall
    Alias for remove
remove
    Remove one or more Jupyter kernelspecs by name.

Comments

1

In Jupyter Notebook, run:

!echo y | jupyter kernelspec uninstall unwanted-kernel

In the Anaconda prompt, run:

jupyter kernelspec uninstall unwanted-kernel

Comments

0

If you are doing this for virtualenv, the kernels in inactive environments might not be shown with jupyter kernelspec list, as suggested above. You can delete it from directory:

~/.local/share/jupyter/kernels/

Comments

0

For Example: To reverse the command python -m ipykernel install --user --name tensorflow --display-name "Python 3.6 (tensorflow)", you need to uninstall the IPython kernel with the same name and display name.

You can do this by running the following command:

jupyter kernelspec uninstall tensorflow

This will remove the IPython kernel with the name "tensorflow" and the display name "Python 3.6 (tensorflow)" from your system.

Comments

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.