I’d expect it to be a fairly common problem when installing a lot of python packages that there would be dependency collisions like package A depending on a certain version of package C and package B depending on another version of package C as a result of which both A and B cannot coexist in a project.
In my case this happens with tensorflow requiring numpy 1.19.2 and matplotlib requiring numpy 1.23.0.
Are there known workarounds to this or do you just have to pick one of them?
Advertisement
Answer
I believe you can try installing an older matplotlib version, which will likely be compatible with the python version required by tensorflow.
That said, I recommend that you use Python’s virtual environment (if you’re not already using it). You will be able to install/test different libraries versions without messing up the ones on your system.