使用CUDA的Tensorflow：ImportError

Question

I've installed TensorFlow step by step like described in this tutorial from NVIDIA (Ubuntu 16.04 Desktop, GTX 970):

http://www.nvidia.com/object/gpu-accelerated-applications-tensorflow-installation.html

Every steps completes without errors, but if I try to import TensorFlow in Python afterwards I'm always gettin' this error message:

>>> import tensorflow as tf
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/__init__.py", line 24, in <module>
    from tensorflow.python import *
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/__init__.py", line 72, in <module>
    raise ImportError(msg)
ImportError: Traceback (most recent call last):
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/__init__.py", line 61, in <module>
    from tensorflow.python import pywrap_tensorflow
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/pywrap_tensorflow.py", line 28, in <module>
    _pywrap_tensorflow = swig_import_helper()
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/pywrap_tensorflow.py", line 24, in swig_import_helper
    _mod = imp.load_module('_pywrap_tensorflow', fp, pathname, description)
ImportError: libcudart.so.8.0: cannot open shared object file: No such file or directory


Failed to load the native TensorFlow runtime.

See https://github.com/tensorflow/tensorflow/blob/master/tensorflow/g3doc/get_started/os_setup.md#import_error

for some common reasons and solutions.  Include the entire stack trace
above this error message when asking for help.

I've installed the CUDA-Toolkit to /opt/cuda/toolkit (also cuDNN). I already tried to symlink:

sudo ln -s /opt/cuda-toolkit /usr/local/cuda
sudo ln -s /opt/cuda-toolkit /usr/local/cuda-8.0

Or this:

sudo export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:/opt/cuda-toolkit/lib64:/opt/cuda-toolkit/extras/CUPTI/lib64
sudo export CUDA_HOME=/opt/cuda-toolkit

But the error remains. Can anybody give me a hint why this error occurs? Thanks a lot :)

Answer 1

I had a similar problem on a machine with similar hardware (GTX 970M) and I solved it using the steps provided in the following post: https://askubuntu.com/a/278840

The only difference in my case was that in the nvidia.conf file I added only the following line: /usr/local/cuda-8.0/lib64

and then ran: sudo ldconfig

Hope it helps :)