Tensor Stream: undefined symbol: cudnnSetRNNDescriptor_v6

Question

when I finished installing shadoworflow (GPU_support, linux 14.04, python3.4) with a virtual environment, according to the instructions of the official site, I checked the installation using the command: python; import of tensor flow; but there is an error:

• import tensorflow as tf Traceback (most recent call last): File "/home/fangfang/tensorflow/lib/python3.4/site-packages/tensorflow/python/pywrap_tensorflow.py", line 58, in <module> from tensorflow.python.pywrap_tensorflow_internal import * File "/home/fangfang/tensorflow/lib/python3.4/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 28, in <module> _pywrap_tensorflow_internal = swig_import_helper() File "/home/fangfang/tensorflow/lib/python3.4/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 "/home/fangfang/tensorflow/lib/python3.4/imp.py", line 243, in load_module return load_dynamic(name, filename, file) ***ImportError: /home/fangfang/tensorflow/lib/python3.4/site-packages/tensorflow/python/../libtensorflow_framework.so: undefined symbol: cudnnSetRNNDescriptor_v6 ***

Answer 1

I had the same error and hopefully this solution will work for you ...

What version of CuDNN are you using? I found that versions tensorflow-gpu> 1.2 would not load until I installed CuDNN v5.1.10.

, , Tensorflow docs , CuDNN v6. https://www.tensorflow.org/install/install_linux#nvidia_requirements_to_run_tensorflow_with_gpu_support

, CuDNN:

$ sudo rm /usr/local/cuda/include/cudnn.h

$ sudo rm /usr/local/cuda/lib64/libcudnn.so

$ sudo rm /usr/local/cuda/lib64/libcudnn.so.5.1.10

v5.1.10. v6 .

CuDNN:

$ cat /usr/local/cuda/include/cudnn.h | grep CUDNN_MAJOR -A 2