0

Has anyone been able to make tensorflow detect the GPU using python 3.7?

How did you do it? I've downloaded cuDNN 8.1, CUDA 11.2, then pip installed tensorflow using pip install tensorflow-gpu==2.5 I've added another environment variable for cuDNN's bin, however I am still getting this result Num GPUs Available 0. Does Tensorflow (CUDA 11.2) even work with the AMD Radeon Vega 8?

Mohan Radhakrishnan
  • 3,002
  • 5
  • 28
  • 42
Xvx
  • 67
  • 1
  • 11
  • Is this mainly about TensorFlow GPU support using AMD hardware ? – Mohan Radhakrishnan Sep 12 '21 at 03:23
  • Relevant link: https://stackoverflow.com/questions/12828268/is-it-possible-to-run-cuda-on-amd-gpus – Jay Mody Sep 12 '21 at 03:28
  • CUDA is an NVIDIA proprietary NVIDIA technology. It doesn't and can't run on AMD GPUs. There is a rocm based port of some parts to Tensorflow to run on AMD GPUs. That isn't distributed through the normal Tensforlow binary release channels AFAIK – talonmies Sep 12 '21 at 05:37
  • @MohanRadhakrishnan yes. Jay Mody Thank you. talonmies Got it, thank you, but do you have an idea where they might have mention that they don't run on AMD GPUs? – Xvx Sep 13 '21 at 02:51

1 Answers1

0

No it does not, because cuDNN is a product of NVIDIA and so is CUDA. NVIDIA designs their own GPUs and their product will look for those GPUs. In order for tensorflow to detect the GPU you will have to use one of NVIDIA's GPU.

Xvx
  • 67
  • 1
  • 11