🐍Python Clean Install
You can follow it if you want to, but it's not finished unless if it's required.
Last updated
You can follow it if you want to, but it's not finished unless if it's required.
Last updated
Alright, go to the Python Page and you should appear in a main page:
Now in this main page, you gonna need is to download version 3.7.X and ONLY 3.7.X! NOT NEWEST VERSION, IT WON'T WORK IF YOU INSTALL IT AND YOU MIGHT GET ERRORS!
Alright now on this page go click downloads and select All releases (I REPEAT DO NOT CLICK 3.10.0 OR YOU MAY GET ERRORS):
After that scroll down till you find 3.7.0 or 3.7.1
Now that you clicked on it, you'll be redirected to this page:
Don't worry, it just shows some fixes and stuff, you should be able to scroll all the way down till you find downloads:
Now you may wanna select "Windows x86-64 executable installer", go ahead and select one and it'll download for you.
Now once it's finished downloading go ahead and open it and you will be greeted by this window:
One thing you should be doing is that you should check "Add Python 3.7 to PATH". This will help CMD to work properly with python.
Go ahead click install or if you want, you can customize installation.
Now wait till it's installed. And once that's done installing you should now go ahead and open the terminal or CMD precise.
(I'll be using Windows Terminal but you can use ether CMD or Powershell)
So you maybe asking, why should we use CMD or Powershell? so that we could test python commands.
So go ahead and type py:
If it happens to be a different version that means you installed incorrect version, if you want you can uninstall that and clear path of it or do the Miniconda setup.
Once you get the right version then congratulations.
Now what you need install is Pytorch and some dependencies.
First off you need to check which CUDA version you have:
If you have CUDA Version 10.2 to 11.3, select CUDA 10.2
Otherwise if you have CUDA 11.3 or up then choose that one.
Other than that, you cannot install PyTorch if you have lower version. (sorry)
For myself I choose CUDA 11.3 so, select that one and copy the entire command, run inside terminal and press enter. Now this may take a while, so be patient...