Issue with installation, suspected cuDNN problem

Hi there,

I am in the process of installing tensorflow through conda environment. So far, I’ve successfully installed Visual Studio, NVIDIA TOOLKIT and NVIDIA cuDNN.
When in conda command prompt after activating my environment if I run the command nvcc --version, an expected response is given (returns the version 11.2 which I installed). However, when I run nvidia-smi, I get " ‘nvidia-smi’ is not recognized as the name of acmdlet, function, script file, or operable program"

In terms of my installation of cuDNN, I downloaded the package and then copied over the bin, include and lib folders in the v11.2 folder of CUDA of NVIDIA Computing Toolkit folder. I installed the 8.1 version, and created my environment with python 3.8. It’s the same error after having uninstalled and reinstalled everything. Please help?


Welcome to the Tensorflow Forum!

Have you followed the steps as described here?

Thank you!