0

I've been trying to install CONDA with Cuda in Centos 7. I am following this installation. When I go through with the install I get no errors thrown until I try to import tensorflow in the environment. When I do, the following stack-trace accompanies a failure to import:

  File "/root/anaconda3/envs/tf_gpu/lib/python3.7/site-packages/tensorflow/python/pywrap_tensorflow.py", line 58, in <module>
    from tensorflow.python.pywrap_tensorflow_internal import *
  File "/root/anaconda3/envs/tf_gpu/lib/python3.7/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 28, in <module>
    _pywrap_tensorflow_internal = swig_import_helper()
  File "/root/anaconda3/envs/tf_gpu/lib/python3.7/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 24, in swig_import_helper
    _mod = imp.load_module('_pywrap_tensorflow_internal', fp, pathname, description)
  File "/root/anaconda3/envs/tf_gpu/lib/python3.7/imp.py", line 242, in load_module
    return load_dynamic(name, filename, file)
  File "/root/anaconda3/envs/tf_gpu/lib/python3.7/imp.py", line 342, in load_dynamic
    return _load(spec)
ImportError: libcuda.so.1: cannot open shared object file: No such file or directory

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/root/anaconda3/envs/tf_gpu/lib/python3.7/site-packages/tensorflow/__init__.py", line 24, in <module>
    from tensorflow.python import pywrap_tensorflow  # pylint: disable=unused-import
  File "/root/anaconda3/envs/tf_gpu/lib/python3.7/site-packages/tensorflow/python/__init__.py", line 49, in <module>
    from tensorflow.python import pywrap_tensorflow
  File "/root/anaconda3/envs/tf_gpu/lib/python3.7/site-packages/tensorflow/python/pywrap_tensorflow.py", line 74, in <module>
    raise ImportError(msg)
ImportError: Traceback (most recent call last):
  File "/root/anaconda3/envs/tf_gpu/lib/python3.7/site-packages/tensorflow/python/pywrap_tensorflow.py", line 58, in <module>
    from tensorflow.python.pywrap_tensorflow_internal import *
  File "/root/anaconda3/envs/tf_gpu/lib/python3.7/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 28, in <module>
    _pywrap_tensorflow_internal = swig_import_helper()
  File "/root/anaconda3/envs/tf_gpu/lib/python3.7/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 24, in swig_import_helper
    _mod = imp.load_module('_pywrap_tensorflow_internal', fp, pathname, description)
  File "/root/anaconda3/envs/tf_gpu/lib/python3.7/imp.py", line 242, in load_module
    return load_dynamic(name, filename, file)
  File "/root/anaconda3/envs/tf_gpu/lib/python3.7/imp.py", line 342, in load_dynamic
    return _load(spec)
**ImportError: libcuda.so.1: 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.

The error is the inability to access the libcuda.so.1 resource. Am I incorrect in thinking that Conda handles all aspects of the installation?

System details:

OS: Centos7

Kernel Version: 3.10.0

Cuda version: 10.0

Conda version: 4.6.14

GPU: Nvidia 1080Ti

  • You need the lib64 directory of your cuda installation added to LD_LIBRARY_PATH environment variable. – Ketan Maheshwari Jun 01 '19 at 23:46
  • Great, thanks. Can you point me to some documentation for that? –  Jun 02 '19 at 00:31
  • @mkc can you provide further clarification? Which path do I export this to? –  Jun 07 '19 at 15:09
  • Look up the web on how to export environment variables and LD_LIBRARY_PATH; then find where is the lib64 directory of your cuda installation. Then put that directory's path to LD_LIBRARY_PATH and export it. – Ketan Maheshwari Jun 12 '19 at 14:39

0 Answers0