TensorFlow GPU版本号与CUDA的对应产生的错误

前言

感悟：cuda 8.0+cudnn 6.0+TensorFlow 1.3 

cuda 9.0+cudnn 7.0+TensorFlow 1.7

python3.6.2+cuda 9.0+cudnn7.5+Tensorflow 1.10.0+Anaconda4.6.11

最近在新的工作站上重新装TensorFlow的GPU版本，刚开始由于省事，直接更新到最新版本1.13，然后输入hello TensorFlow程序。但是却报错“ImportError: DLL load failed: 找不到指定的模块”。无奈之下，各种百度，看到有个比较旧博客提议将TensorFlow版本降低到1.4，于是先卸载再重装，一顿修改之后，又报错“Could not find 'cudart64_80.dll'. TensorFlow requires that this DLL be installed in a directory that is named in your %PATH% environment variable. Download and install CUDA 8.0 from this URL: https://developer.nvidia.com/cuda-toolkit”，这句话的意思就是说我装的TensorFlow版本太低，只能支持CUDA8.0，但是我装的是CUDA9.0，所以出现了不对应。后来，又卸载当前TensorFlow环境，指定安装1.7版本，搞定。特此记录下来，防止后人少踩坑。

以下图示均为命令行操作

TensorFlow版本过低，CUDA版本过高

 具体报错如下：

(tensorflow-gpu) C:UsersWW>python
Python 3.6.2 |Continuum Analytics, Inc.| (default, Jul 20 2017, 12:30:02) [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 "D:TensorFlowAnacondaAnacondaenvs	ensorflow-gpulibsite-packages	ensorflowpythonplatformself_check.py", line 75, in preload_check
    ctypes.WinDLL(build_info.cudart_dll_name)
  File "D:TensorFlowAnacondaAnacondaenvs	ensorflow-gpulibctypes\__init__.py", line 348, in __init__
    self._handle = _dlopen(self._name, mode)
OSError: [WinError 126] 找不到指定的模块。

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "D:TensorFlowAnacondaAnacondaenvs	ensorflow-gpulibsite-packages	ensorflow\__init__.py", line 24, in <module>
    from tensorflow.python import *
  File "D:TensorFlowAnacondaAnacondaenvs	ensorflow-gpulibsite-packages	ensorflowpython\__init__.py", line 49, in <module>
    from tensorflow.python import pywrap_tensorflow
  File "D:TensorFlowAnacondaAnacondaenvs	ensorflow-gpulibsite-packages	ensorflowpythonpywrap_tensorflow.py", line 30, in <module>
    self_check.preload_check()
  File "D:TensorFlowAnacondaAnacondaenvs	ensorflow-gpulibsite-packages	ensorflowpythonplatformself_check.py", line 82, in preload_check
    % (build_info.cudart_dll_name, build_info.cuda_version_number))
ImportError: Could not find 'cudart64_80.dll'. TensorFlow requires that this DLL be installed in a directory that is named in your %PATH% environment variable. Download and install CUDA 8.0 from this URL: https://developer.nvidia.com/cuda-toolkit

TensorFlow版本过高，CUDA版本过低

 具体错误如下所示：

(tensorflow-gpu) C:UsersWW>python
Python 3.6.2 |Continuum Analytics, Inc.| (default, Jul 20 2017, 12:30:02) [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 "D:TensorFlowAnacondaAnacondaenvs	ensorflow-gpulibsite-packages	ensorflowpythonpywrap_tensorflow.py", line 58, in <module>
    from tensorflow.python.pywrap_tensorflow_internal import *
  File "D:TensorFlowAnacondaAnacondaenvs	ensorflow-gpulibsite-packages	ensorflowpythonpywrap_tensorflow_internal.py", line 28, in <module>
    _pywrap_tensorflow_internal = swig_import_helper()
  File "D:TensorFlowAnacondaAnacondaenvs	ensorflow-gpulibsite-packages	ensorflowpythonpywrap_tensorflow_internal.py", line 24, in swig_import_helper
    _mod = imp.load_module('_pywrap_tensorflow_internal', fp, pathname, description)
  File "D:TensorFlowAnacondaAnacondaenvs	ensorflow-gpulibimp.py", line 242, in load_module
    return load_dynamic(name, filename, file)
  File "D:TensorFlowAnacondaAnacondaenvs	ensorflow-gpulibimp.py", line 342, in load_dynamic
    return _load(spec)
ImportError: DLL load failed: 找不到指定的模块。

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "D:TensorFlowAnacondaAnacondaenvs	ensorflow-gpulibsite-packages	ensorflow\__init__.py", line 24, in <module>
    from tensorflow.python import pywrap_tensorflow  # pylint: disable=unused-import
  File "D:TensorFlowAnacondaAnacondaenvs	ensorflow-gpulibsite-packages	ensorflowpython\__init__.py", line 49, in <module>
    from tensorflow.python import pywrap_tensorflow
  File "D:TensorFlowAnacondaAnacondaenvs	ensorflow-gpulibsite-packages	ensorflowpythonpywrap_tensorflow.py", line 74, in <module>
    raise ImportError(msg)
ImportError: Traceback (most recent call last):
  File "D:TensorFlowAnacondaAnacondaenvs	ensorflow-gpulibsite-packages	ensorflowpythonpywrap_tensorflow.py", line 58, in <module>
    from tensorflow.python.pywrap_tensorflow_internal import *
  File "D:TensorFlowAnacondaAnacondaenvs	ensorflow-gpulibsite-packages	ensorflowpythonpywrap_tensorflow_internal.py", line 28, in <module>
    _pywrap_tensorflow_internal = swig_import_helper()
  File "D:TensorFlowAnacondaAnacondaenvs	ensorflow-gpulibsite-packages	ensorflowpythonpywrap_tensorflow_internal.py", line 24, in swig_import_helper
    _mod = imp.load_module('_pywrap_tensorflow_internal', fp, pathname, description)
  File "D:TensorFlowAnacondaAnacondaenvs	ensorflow-gpulibimp.py", line 242, in load_module
    return load_dynamic(name, filename, file)
  File "D:TensorFlowAnacondaAnacondaenvs	ensorflow-gpulibimp.py", line 342, in load_dynamic
    return _load(spec)
ImportError: DLL load failed: 找不到指定的模块。


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.

TensorFlow与CUDA版本的对应关系

附上几张表格：

具体最新版本对应可参考TensorFlow中文网址：https://www.tensorflow.org/install/source#tested_source_configurations

总结

1. 安装环境时参考的博客一定要注意时间，时间，时间。有可能当时可以的现在就不一定行了，版本问题真的很烦人呐呐呐
2. 切勿贪图省事，更新到最新版本，要提前了解清楚，然后再装对应的版本

参考

https://blog.csdn.net/yeler082/article/details/80943040