I have installed Cuda 10.1 and cudnn on Ubuntu 18.04 and it seems to be installed properly as type nvcc and nvidia-smi, I get proper response:
user:~$ nvcc -V
I had the correct version of CUDA and tensorflow-gpu==1.14.0
installed on my conda environment, but somehow I was still getting this error message. This post helped me to finally solve it.
I had previously installed tensorflow-gpu
via pip
- after creating a new environment and installing tensorflow-gpu
via conda
solved my problem.
conda install -c anaconda tensorflow-gpu=1.14.0