Installing Tensorflow 2.0, I have GPU but not CUDA-Enabled, Does that means I can only used my CPU?

0
votes

I read the TensorFlow documentation about how to install Tensorflow 2.0 with GPU on Windows (https://www.tensorflow.org/install/gpu).

In the hardware requirement section, it states that it needs NVIDIA GPU with CUDA Compute Capability. When I saw CUDA-enabled GPU cards list (https://developer.nvidia.com/cuda-gpus), I just realized that my GPU was not on the list (I have NVIDIA GTX 1660 ti on my machine).

What I wanted to ask are:

  1. Does it mean that I can't use my GPU for my Deep Learning Tensorflow implementation?
  2. Does it mean that I have to install the CPU version of Tensorflow 2.0?
1
pythontensorflow
Your GPU has full CUDA support – talonmies
but why I can't find GTX1660ti in here developer.nvidia.com/cuda-gpus ? TensorFlow documentation suggests that I should check my GPU on that list first. – Dennis
Because someone at Nvidia didn't update the list – talonmies
I hope so, thx for the info! – Dennis

1 Answers

1
votes

You can use Tensorflow GPU on your laptop as your GPU supports CUDA. Install CUDA, then cuDNN and then, Tensorflow GPU.

You don't have to work with CPU version as you have a better option.