Import error: cannot open shared object file:no such file or directory

Question

Problem :

I have installed Cuda 10.1 and cudnn on Ubuntu 18.04 for the first time and it seems to be installed properly as when I type nvcc and nvidia-smi I am getting proper response.

I have also installed tensorflow using following command:

user:~$ sudo pip3 install --upgrade tensorflow-gpu

But when I am trying to import tensorflow I am getting following error about libcublas.so.10.0:

ImportError : libcublas.so.10.0: cannot open shared object file: No such file or directory

Failed to load the native TensorFlow runtime.

See https://www.tensorflow.org/install/errors

for some common reasons and solutions.  Include the entire stack trace

above this error message when asking for help.

Am I missing something? How to resolve this issue?

Answer 1

This error generally occurs when the version of your cuda and tensorflow installed are not compatible. I also encountered the same Import Error while I was running tensorflow version 1.13.0 with my cuda 9. As I had installed tensorflow on a virtual environment with pip, I just had to uninstall the tensorflow 1.13.0 and install tensorflow 1.12.0 using following commands :

pip uninstall tensorflow-gpu tensorflow-estimator tensorboard

pip install tensorflow-gpu==1.12.0

Now everything is working for me.

OR

You can fix the error by adding the following command to your '.bashrc' file.

export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda-10.0/lib64/

 

System configuration:

Ubuntu :- 16.04 LTS

Tensorflow :-GPU 2.0beta1

Cuda:- 10.0

cuDNN :-7.6.0 for Cuda 10.0

Here I used the conda to configure my system.