Hello guys, new user here.
Using Tensorflow on Windows with CUDA became much more complicated after it dropped native support for Windows. Before I head dive into it, let me understand the situation:
- If I want to use CUDA GPU easily, maybe I should just stick to Tensorflow 2.10 for now?
- I am not experienced with this WSL2 thing and can’t quite see how does it interact with Windows. If I build and install Tensorflow in WSL2, will I be able to use GPU from scripts launched on Windows? If not, what is the purpose of all of that?
Thanks for any insight.