Tensorflow-gpu installation errors in windows

Question

Recently I was installing tensorflow(gpu) for windows. I have a graphic card with compute capability = 3.0 (compatible) Python (3.5.2) Cuda 8.0 (installed and working) CudNN installed (These are as per the tensorflow manual) But when I installed the tensorflow

pip3 install --upgrade tensorflow-gpu

It was installed successfully. But when I run the code in python I got the error. Here is the log.

>>C:\Users\jagja>python
Python 3.5.2 (v3.5.2:4def2a2901a5, Jun 25 2016, 22:18:55) [MSC v.1900 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import tensorflow
Traceback (most recent call last):
  File "C:\Users\jagja\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\pywrap_tensorflow_internal.py", line 18, in swig_import_helper
    return importlib.import_module(mname)
  File "C:\Users\jagja\AppData\Local\Programs\Python\Python35\lib\importlib\__init__.py", line 126, in import_module
    return _bootstrap._gcd_import(name[level:], package, level)
  File "<frozen importlib._bootstrap>", line 986, in _gcd_import
  File "<frozen importlib._bootstrap>", line 969, in _find_and_load
  File "<frozen importlib._bootstrap>", line 958, in _find_and_load_unlocked
  File "<frozen importlib._bootstrap>", line 666, in _load_unlocked
  File "<frozen importlib._bootstrap>", line 577, in module_from_spec
  File "<frozen importlib._bootstrap_external>", line 906, in create_module
  File "<frozen importlib._bootstrap>", line 222, in _call_with_frames_removed
ImportError: DLL load failed: The specified module could not be found.

>>During handling of the above exception, another exception occurred:

>>Traceback (most recent call last):
  File "C:\Users\jagja\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\pywrap_tensorflow.py", line 41, in <module>
    from tensorflow.python.pywrap_tensorflow_internal import *
  File "C:\Users\jagja\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\pywrap_tensorflow_internal.py", line 21, in <module>
    _pywrap_tensorflow_internal = swig_import_helper()
  File "C:\Users\jagja\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\pywrap_tensorflow_internal.py", line 20, in swig_import_helper
    return importlib.import_module('_pywrap_tensorflow_internal')
  File "C:\Users\jagja\AppData\Local\Programs\Python\Python35\lib\importlib\__init__.py", line 126, in import_module
    return _bootstrap._gcd_import(name[level:], package, level)
ImportError: No module named '_pywrap_tensorflow_internal'

>>During handling of the above exception, another exception occurred:

>>Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "C:\Users\jagja\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\__init__.py", line 24, in <module>
    from tensorflow.python import *
  File "C:\Users\jagja\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\__init__.py", line 49, in <module>
    from tensorflow.python import pywrap_tensorflow
  File "C:\Users\jagja\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\pywrap_tensorflow.py", line 52, in <module>
    raise ImportError(msg)
ImportError: Traceback (most recent call last):
  File "C:\Users\jagja\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\pywrap_tensorflow_internal.py", line 18, in swig_import_helper
    return importlib.import_module(mname)
  File "C:\Users\jagja\AppData\Local\Programs\Python\Python35\lib\importlib\__init__.py", line 126, in import_module
    return _bootstrap._gcd_import(name[level:], package, level)
  File "<frozen importlib._bootstrap>", line 986, in _gcd_import
  File "<frozen importlib._bootstrap>", line 969, in _find_and_load
  File "<frozen importlib._bootstrap>", line 958, in _find_and_load_unlocked
  File "<frozen importlib._bootstrap>", line 666, in _load_unlocked
  File "<frozen importlib._bootstrap>", line 577, in module_from_spec
  File "<frozen importlib._bootstrap_external>", line 906, in create_module
  File "<frozen importlib._bootstrap>", line 222, in _call_with_frames_removed
ImportError: DLL load failed: The specified module could not be found.

>>During handling of the above exception, another exception occurred:

>>Traceback (most recent call last):
  File "C:\Users\jagja\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\pywrap_tensorflow.py", line 41, in <module>
    from tensorflow.python.pywrap_tensorflow_internal import *
  File "C:\Users\jagja\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\pywrap_tensorflow_internal.py", line 21, in <module>
    _pywrap_tensorflow_internal = swig_import_helper()
  File "C:\Users\jagja\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\pywrap_tensorflow_internal.py", line 20, in swig_import_helper
    return importlib.import_module('_pywrap_tensorflow_internal')
  File "C:\Users\jagja\AppData\Local\Programs\Python\Python35\lib\importlib\__init__.py", line 126, in import_module
    return _bootstrap._gcd_import(name[level:], package, level)
ImportError: No module named '_pywrap_tensorflow_internal'


>>Failed to load the native TensorFlow runtime.

>>See https://www.tensorflow.org/install/install_sources#common_installation_problems

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

Can anyone figure out the problem. I have trying to work upon gpu's now, but stuck for weeks because of this error

Answer 1

I tried really hard to find a solution to this problem problem. I don't know which one is necessary or not for you. But lastly I did this and it worked:

You must change cudnn64_?.dll in directory C:\\Program Files\\NVIDIA GPU Computing Toolkit\\CUDA\\v8.0\\bin to cudnn64_5.dll .

I hope it work for you too.

Update In fact you must download adequate version of CuDNN for your tensorflow. In windows 10 version 5.1 works.

Answer 2

Setting up the development environment for GPU computations for deep learning/ML libraries is extremely frustrating. In my experience, it helps often to start from a blank slate again if the error messages get too convoluted. But this particular error message seems to be one of the common installation problems mentioned on tensorflow's installation page. They redirect to this stackoverflow question :

On Windows, running "import tensorflow" generates No module named "_pywrap_tensorflow" error

It is extremely hard to diagnose something like this from afar. If the link doesnt help, I hope someone else can come along and help you. Good luck!