Tensorflow 2.0 list_physical_devices doesn't detect my GPU

Question

I recently install tensorflow 2.0 on my computer but when I try to run it on my GPU, the function tf.config.experimental.list_physical_devices('GPU') on Jupyter or Vitual Studio Code it returns me a void array. Do you know why?

My set-up:

Computer: MSI

Processor: Intel(R) Core(TM) i7-8750H CPU @ 2.220GHz

GPU 0: Intel(R) UHD Graphics 630

GPU: NVIDIA GeForce GTX 1060

Python: Ananconda 3 with Python 3.7

Tensenflow 2.0 installed with pip install tensorflow

My test code:

physical_devices = tf.config.experimental.list_physical_devices('GPU')
print(physical_devices)
if physical_devices:
  tf.config.experimental.set_memory_growth(physical_devices[0], True)

Thanks in advance: :)

Answer 1

Providing the solution here (Answer Section), even though it is present in the Comment Section for the benefit of the community.

Instead of pip install tensorflow , you can try pip3 install --upgrade tensorflow-gpu or just remove tensorflow and then installing "tensorflow-gpu will resolves your issue.

After installation of Tensorflow GPU, you can check GPU as below

physical_devices = tf.config.experimental.list_physical_devices('GPU')
print(physical_devices)
if physical_devices:
  tf.config.experimental.set_memory_growth(physical_devices[0], True)

Output:

[PhysicalDevice(name='/physical_device:GPU:0', device_type='GPU')]

Answer 2

Upgrading simply worked for me.:

pip3 install --upgrade tensorflow-gpu

and the device name searched has to be 'XLA_GPU', and does not respond with solo 'GPU' search term. But it also raised another error when setting the memory growth which is not supported by 'XLA' gpus.

Answer 3

Installing tensorflow with gpu using Conda

First install anaconda or mini-anaconda on your system and make sure you have set the environment path for conda command.

In below command replace tensor with a environment name of your choice:

conda create -n tensor tensorflow-gpu cudatoolkit=9.0
conda activate tensor

Then update your tensorflow-gup package using pip and make sure your environment is activated in conda

pip3 install --upgrade tensorflow-gpu

Verify weather tensorflow with gpu support is installed successfully

(tensor) C:\> python

Python 3.8.15 (default, Nov 24 2022, 14:38:14) [MSC v.1916 64 bit (AMD64)] :: Anaconda, Inc. on win32
Type "help", "copyright", "credits" or "license" for more information.

>>> import tensorflow
>>> tensorflow.config.list_physical_devices("GPU")
[PhysicalDevice(name='/physical_device:GPU:0', device_type='GPU')]

The output [PhysicalDevice(name='/physical_device:GPU:0', device_type='GPU')] in terminal means we have successfully installed tensorflow with gpu.

This answer was taken fromanaconda tensorflow installation link.