I’m not able to resolve a CUDA dependency to run tensorflow models. I’m not sure what installation step I’m missing.

Following is what I have on my Windows 8 system:

On nvcc -version

Cuda compilation tools, release 11.1, V11.1.74
Build cuda_11.1.relgpu_drvr455TC455_06.29069683_0

On conda list cudatoolkit

# Name                    Version                   Build  Channel
cudatoolkit               10.2.89              h74a9793_1

Tensorflow version 2.4.0-dev20200928

Python Version 3.8.3

tf-nightly-gpu is latest

I’m getting the following errors:

2020-09-30 15:25:07.471899: W tensorflow/stream_executor/platform/default/dso_loader.cc:59] Could not load dynamic library 'cudnn64_8.dll'; dlerror: cudnn64_8.dll not found
2020-09-30 15:25:07.472033: W tensorflow/core/common_runtime/gpu/gpu_device.cc:1753] Cannot dlopen some GPU libraries. Please make sure the missing libraries mentioned above are installed properly if you would like to use GPU. Follow the guide at https://www.tensorflow.org/install/gpu for how to download and setup the required libraries for your platform.
Skipping registering GPU devices...

Source: Python Questions