Greetings.
I am Khan Mohammad Azam from Korea University 'DAVIAN' team. We found the following error in 'stderr' file after finishing the job.
Job ID: 4439ae43-9b37-4d7c-b704-f28935dbcad9
...
Using TensorFlow backend.
Traceback (most recent call last):
File "/usr/local/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow.py", line 58, in
from tensorflow.python.pywrap_tensorflow_internal import *
File "/usr/local/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 28, in
_pywrap_tensorflow_internal = swig_import_helper()
File "/usr/local/lib/python3.6/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 "/usr/local/lib/python3.6/imp.py", line 243, in load_module
return load_dynamic(name, filename, file)
File "/usr/local/lib/python3.6/imp.py", line 343, in load_dynamic
return _load(spec)
ImportError: libcublas.so.9.0: cannot open shared object file: No such file or directory
...
...
Should we make the following export or similar one somewhere in script?
**export LD_LIBRARY_PATH=LD_LIBRARY_PATH:/usr/local/cuda-9.0/lib64/**
I am just wondering if you can notify us whether the problem is due to the above reason or someting else. Thank you very much.
Best,
KU DAVIAN TEAM.
Created by ku.davian.lab Thank you for your kind and valuable directions. Yes, you are correct, it was a base image issue that was created by me. It is resolved now. This is capture from our task server.
```
$ docker run --runtime=nvidia -e NVIDIA_VISIBLE_DEVICES=7 -ti tensorflow/tensorflow:1.12.0-gpu-py3 /bin/bash
root@895b5d69c7aa:/notebooks# python
Python 3.5.2 (default, Nov 23 2017, 16:37:01)
[GCC 5.4.0 20160609] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import tensorflow as tf
>>> tf.__version__
'1.12.0'
>>>
root@895b5d69c7aa:/notebooks# nvidia-smi
Mon Dec 17 04:09:13 2018
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 390.48 Driver Version: 390.48 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 Tesla P40 On | 00000000:22:00.0 Off | 0 |
| N/A 18C P8 11W / 250W | 0MiB / 22919MiB | 0% Default |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| No running processes found |
+-----------------------------------------------------------------------------+
root@895b5d69c7aa:/notebooks# exit
exit
```
Your issue is obviously a base image issue.
Thank you for your kind suggestions. Actually, I have tried several images including the one you have mentioned. Perhaps this is an issue to share GPU drivers from host to containers.
I am just wondering if any other team successfully resolve this issue.
When I run the following command inside the container, it gives the above error as well.
$ python
>> import tensorflow as tf
...
...
ImportError: libcublas.so.9.0: cannot open shared object file: No such file or directory
...
...
We appreciate any suggestion to overcome this issue. Thank you. Would you like to try this image 'tensorflow/tensorflow:1.12.0-gpu-py3' ?
Or search the environments you want from 'https://hub.docker.com/r/tensorflow/tensorflow'.
Drop files to upload
ImportError: libcublas.so.9.0: cannot open shared object file: No such file or directory page is loading…