(In conda environment) The base environment uses python3.8, but as of 2020/09/22, tensorflow does not support python3.8. Therefore
--Create an environment called py37 for python3.7 --Since it is troublesome to reinstall all the packages, take over from base or something (this time take over from data analysis environment ds)
I want to do something like that.
conda create -n py37 python=3.7
Now you can create an environment called py37 for python3.7
Next, I want to put the packages used in ds in this environment. First, output the ds package list as txt.
conda list --explicit > ds.txt
ds.txt is created in the current working directory.
Since ds used python3.8, there is a place to install python3.8 in ds.txt. Specifically, the following line
https://repo.anaconda.com/pkgs/main/win-64/python-3.8.5-h5fd99cc_1.conda
Open ds.txt with an editor or something and delete this line.
Reflect the contents of ds.txt in py37.
conda activate py37
After activating py37 with
conda update --file ds.txt
With this, the newly created py37 will inherit the ds environment other than the python version.
https://repo.anaconda.com/pkgs/main/win-64/python-3.8.5-h5fd99cc_1.conda
If this sentence is properly removed from ds.txt
conda create -n py37 python=3.7 --file ds.txt
So, I think I can create and take over the environment, but I haven't tried it.
It's a fairly manual approach, so please comment if there's an easier way to do it.
Recommended Posts