3

I am trying to follow this post to change the name of my kernels in Jupyter to make more sense, however jupyter kernelspec list doesn't find all the kernels.

(base) bash-3.2$ which python
/Users/benjamin/anaconda3/bin/python
(base) bash-3.2$ which jupyter
/Users/benjamin/anaconda3/bin/jupyter
(base) bash-3.2$ jupyter kernelspec list
Available kernels:
  python3    /Users/benjamin/Library/Jupyter/kernels/python3

However in Jupyter I can see more than one Kernel

enter image description here

If I look in a different location I get the same result:

(base) bash-3.2$ cd anaconda3/share/jupyter/kernels/
(base) bash-3.2$ ls
python3

Why is this and how can I fix it?

Bn.F76
  • 783
  • 2
  • 12
  • 30
  • 1
    There's a related open issue on `nb_conda_kernels` repo: [Conda environments not detected](https://github.com/Anaconda-Platform/nb_conda_kernels/issues/112) – merv Jun 04 '19 at 06:09
  • 1
    Update: This is not an extra kernel `jupyter kernelspec list` supposed to list. If you `conda remove nb_conda_kernels` from within the environment, then `Python [conda env:root]` will also disappear from Jupyter. – Bn.F76 Apr 17 '20 at 22:44

1 Answers1

0

I just have a similar problem, not realized before (due to i have over 10 kernels). Finally i sorted out that:

a) If you dont define explicitly a Kernel (name given by you using the "jupyter kernelspec" command, you will see just an Environment name set by the Conda system

b) This "Environment xyz" will be only accessible while you are in the base env, or when you are in xyz env, NOT whilst your are in other envs

c) However if you create an explicit Kernel xyz, then that Kernel xyz is then available in other environments

d) It looks like the name of that Environment xyz follow the name of your environment (if that is the case that you can't really change it, but you can change the associated kernel's name). That could be confusing when you are having a lot of different kernels/environment

r poon
  • 633
  • 7
  • 7