How to install tensorflow-gpu?

Question

How to install tensorflow-gpu on windows 10 with Python 3.10?

conda and pip do not work

I installed:

cudnn-windows-x86_64-8.9.0.131_cuda11-archive
cuda_12.1.1_531.14_windows

add to path

C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.1\bin
C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.1\libnvvp

conda create --name cuda
conda activate cuda


(cuda) C:\Users\xxx>python -V
Python 3.10.10


(cuda) C:\Users\xxx>conda install -c conda-forge tensorflow-gpu
Collecting package metadata (current_repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: failed with repodata from current_repodata.json, will retry with next repodata source.
Collecting package metadata (repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.


(cuda) C:\Users\xxx>pip install -U tensorflow-gpu
Collecting tensorflow-gpu
  Using cached tensorflow-gpu-2.12.0.tar.gz (2.6 kB)
  Preparing metadata (setup.py) ... error
  error: subprocess-exited-with-error

  × python setup.py egg_info did not run successfully.
  │ exit code: 1
  ╰─> [39 lines of output]
      Traceback (most recent call last):
        File "C:\Users\xxx\anaconda3\envs\cuda\lib\site-packages\setuptools\_vendor\packaging\requirements.py", line 35, in __init__
          parsed = parse_requirement(requirement_string)
        File "C:\Users\xxx\anaconda3\envs\cuda\lib\site-packages\setuptools\_vendor\packaging\_parser.py", line 64, in parse_requirement
          return _parse_requirement(Tokenizer(source, rules=DEFAULT_RULES))
        File "C:\Users\xxx\anaconda3\envs\cuda\lib\site-packages\setuptools\_vendor\packaging\_parser.py", line 82, in _parse_requirement
          url, specifier, marker = _parse_requirement_details(tokenizer)
        File "C:\Users\xxx\anaconda3\envs\cuda\lib\site-packages\setuptools\_vendor\packaging\_parser.py", line 126, in _parse_requirement_details
          marker = _parse_requirement_marker(
        File "C:\Users\xxx\anaconda3\envs\cuda\lib\site-packages\setuptools\_vendor\packaging\_parser.py", line 147, in _parse_requirement_marker
          tokenizer.raise_syntax_error(
        File "C:\Users\xxx\anaconda3\envs\cuda\lib\site-packages\setuptools\_vendor\packaging\_tokenizer.py", line 163, in raise_syntax_error
          raise ParserSyntaxError(
      setuptools.extern.packaging._tokenizer.ParserSyntaxError: Expected end or semicolon (after name and no valid version specifier)
          python_version>"3.7"
                        ^

      The above exception was the direct cause of the following exception:

      Traceback (most recent call last):
        File "<string>", line 2, in <module>
        File "<pip-setuptools-caller>", line 34, in <module>
        File "C:\Users\xxx\AppData\Local\Temp\pip-install-t4yr5nvl\tensorflow-gpu_48989e8e399a4c5da19c2d876e93f0d7\setup.py", line 40, in <module>
          setuptools.setup()
        File "C:\Users\xxx\anaconda3\envs\cuda\lib\site-packages\setuptools\__init__.py", line 106, in setup
          _install_setup_requires(attrs)
        File "C:\Users\xxx\anaconda3\envs\cuda\lib\site-packages\setuptools\__init__.py", line 77, in _install_setup_requires
          dist.parse_config_files(ignore_option_errors=True)
        File "C:\Users\xxx\anaconda3\envs\cuda\lib\site-packages\setuptools\dist.py", line 910, in parse_config_files
          self._finalize_requires()
        File "C:\Users\xxx\anaconda3\envs\cuda\lib\site-packages\setuptools\dist.py", line 607, in _finalize_requires
          self._move_install_requirements_markers()
        File "C:\Users\xxx\anaconda3\envs\cuda\lib\site-packages\setuptools\dist.py", line 647, in _move_install_requirements_markers
          inst_reqs = list(_reqs.parse(spec_inst_reqs))
        File "C:\Users\xxx\anaconda3\envs\cuda\lib\site-packages\setuptools\_vendor\packaging\requirements.py", line 37, in __init__
          raise InvalidRequirement(str(e)) from e
      setuptools.extern.packaging.requirements.InvalidRequirement: Expected end or semicolon (after name and no valid version specifier)
          python_version>"3.7"
                        ^
      [end of output]

  note: This error originates from a subprocess, and is likely not a problem with pip.
error: metadata-generation-failed

× Encountered error while generating package metadata.
╰─> See above for output.

note: This is an issue with the package mentioned above, not pip.
hint: See above for details.

i found a solution with pycharm and ubuntu in windows but i don't want to do it.

Answer 1

New Solution (Command Line)

Edit: It is now far easier to download Tensorflow with GPU support using the command line. I have kept the old solution below, but I'd recommend you use this new solution.

For Windows, you'll need to use Conda from the command line.

conda install -c conda-forge cudatoolkit=11.2 cudnn=8.1.0
# Anything above 2.10 is not supported on the GPU on Windows Native
python -m pip install "tensorflow<2.11"
# Verify the installation:
python -c "import tensorflow as tf; print(tf.config.list_physical_devices('GPU'))"

For Linux, you can download using pip.

python3 -m pip install 'tensorflow[and-cuda]'
# Verify the installation:
python3 -c "import tensorflow as tf; print(tf.config.list_physical_devices('GPU'))"

There is no official support for Mac OS

Old Solution

It is now deprecated. Check release notes for more information: https://github.com/tensorflow/tensorflow/releases

Or the pypi page: https://pypi.org/project/tensorflow-gpu

If you're looking for a tutorial, this one worked for me: https://www.youtube.com/watch?v=OEFKlRSd8Ic

One small thing I had to change was the final step of the tutorial. As of August 2023, you have to install like follows:

pip install tensorflow==2.10.0

Then check if it works like this:

>>> import tensorflow as tf
>>> tf.test.is_built_with_cuda()

Answer 2

according to pypi:

tensorflow-gpu has been removed. Please install tensorflow instead. The tensorflow package supports GPU accelerated operations via Nvidia CUDA.

so just use the following command for install:

pip install tensorflow