Windows 11, rtx 4070 error installing cudnn/tensorflow

Accelerated Computing CUDA CUDA Setup and Installation
1

i have cudnn 8.8.1.3 installed as well as cuda 12.1.0_531.14
i’m following the steps from tensorflow on windows-wsl2 :
python -m pip install nvidia-cudnn-cu11==8.6.0.163 tensorflow==2.12.*

gives me an error:
Looking in indexes: https pypi org simple, https pypi ngc nvidia com (edit: link removed, can’t paste it here)
ERROR: Could not find a version that satisfies the requirement nvidia-cudnn-cu11==8.6.0.163 (from versions: 0.0.1.dev5, 2021.10.26, 2021.11.18, 2021.12.8, 2022.1.13, 2022.4.2, 2022.5.19)
ERROR: No matching distribution found for nvidia-cudnn-cu11==8.6.0.163

same if i change version to 8.8.1.3

2 Likes
2

I’m trying to install cudnn version 8.6.0.163 and am getting the same error. The error also happens for other versions. I’m on Windows 11, RTX 3060.

When I run pip install nvidia-cudnn-cu11 I get:

Looking in indexes: https://pypi.org/simple, https://pypi.ngc.nvidia.com
Collecting nvidia-cudnn-cu11
  Downloading https://developer.download.nvidia.com/compute/redist/nvidia-cudnn-cu11/nvidia-cudnn-cu11-2022.5.19.tar.gz (16 kB)
  Preparing metadata (setup.py) ... done
Collecting nvidia-cudnn-cu116
  Downloading nvidia-cudnn-cu116-0.0.1.dev5.tar.gz (4.4 kB)
  Preparing metadata (setup.py) ... error
  error: subprocess-exited-with-error

  × python setup.py egg_info did not run successfully.
  │ exit code: 1
  ╰─> [17 lines of output]
      Traceback (most recent call last):
        File "<string>", line 2, in <module>
        File "<pip-setuptools-caller>", line 34, in <module>
        File "C:\Users\mallo\AppData\Local\Temp\pip-install-b6w1lra5\nvidia-cudnn-cu116_43c376fad8fd4c15b8b5516f5eaa4c32\setup.py", line 152, in <module>
          raise RuntimeError(open("ERROR.txt", "r").read())
      RuntimeError:
      ###########################################################################################
      The package you are trying to install is only a placeholder project on PyPI.org repository.
      This package is hosted on NVIDIA Python Package Index.

      This package can be installed as:
      ```
      $ pip install nvidia-pyindex
      $ pip install nvidia-cudnn-cu116
      ```
      ###########################################################################################

      [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.

Does anyone know how to solve this? Thanks.

2 Likes
3

There is an issue with tensorflow >=2.11. This solved my problem:

Wish you luck!

4

I have the problem too. Even with WSL 2 installed on Windows 11. The step that fails is when trying to install cudnn as described in the TensorFlow docs. But I am not convinced that this is a TensorFlow installation issue.

Command I issued:
pip install nvidia-cudnn-cu11==8.6.0.163
Error:
ERROR: Could not find a version that satisfies the requirement nvidia-cudnn-cu11==8.6.0.163 (from versions: 0.0.1.dev5, 2021.10.26, 2021.11.18, 2021.12.8, 2022.1.13, 2022.4.2, 2022.5.19)
ERROR: No matching distribution found for nvidia-cudnn-cu11==8.6.0.163

This is happening, before TensorFlow is installed. nvidia-cudnn-cu11 is mainTained by nvidia, SEE nvidia-cudnn-cu11 · PyPI

2 Likes
5

same issue, bother me a lot

2 Likes