Hello,
I would like to use tensorflow with a graphic card (I put a A2000 graphic card), but when I try to import it, It says that it can’t regognise (because cuba driver not found), and It uses my cpu when I train my model.
I don’t kknow which cuba driver I’ve to download (there isn’t zimaos disponibility on nvidia website).
Someone used his graphic card to train model with tensorflow? (I use anaconda docker).
Do I need to confirm that your system is ZimaOS? Can you see the graphics card usage in the interface?
Of course it is zimaos, and I can see.
I have some news, I think gpu is not linked to docker anaconda3 (I don’t know how to link it, I searched on each community, try various solutions without success).
I changed the daemons.json so each docker can normally have access (but anaconda still doesn’t find).
Now I used Tensorflow-gpu docker (there is one with jupyter) and now It can use my gpu
I think I will replace anaconda3 with tensorflow-jupyter, for me it is the same.
But if someone can use anaconda3 with his gpu, let me know please!
I apologize for replying to you so late, I understand that you need a Python development environment？