我安装了 Tensorflow gpu 版本和 CUDA 9.1 和 python 3.6.5。但是，在我导入 tensorflow-gpu 后，出现以下错误：答案

【问题标题】：I installed the Tensorflow gpu version and CUDA 9.1 and python 3.6.5. However, after I import the tensorflow-gpu I got the following error:我安装了 Tensorflow gpu 版本和 CUDA 9.1 和 python 3.6.5。但是，在我导入 tensorflow-gpu 后，出现以下错误：
【发布时间】：2018-10-27 00:12:19
【问题描述】：

Traceback（最近一次调用最后一次）：

  File "C:\Users\ranji\AppData\Local\Programs\Python\Python36\lib\site-packages\tensorflow\python\platform\self_check.py", line 75, in preload_check
    ctypes.WinDLL(build_info.cudart_dll_name)
  File "C:\Users\ranji\AppData\Local\Programs\Python\Python36\lib\ctypes\__init__.py", line 348, in __init__
    self._handle = _dlopen(self._name, mode)
OSError: [WinError 126] The specified module could not be found

在处理上述异常的过程中，又发生了一个异常：

Traceback (most recent call last):
  File "C:/Users/ranji/AppData/Local/Programs/Python/Python36/ts1.py", line 1, in <module>
    import tensorflow as tf
  File "C:\Users\ranji\AppData\Local\Programs\Python\Python36\lib\site-packages\tensorflow\__init__.py", line 24, in <module>
    from tensorflow.python import pywrap_tensorflow  # pylint: disable=unused-import
  File "C:\Users\ranji\AppData\Local\Programs\Python\Python36\lib\site-packages\tensorflow\python\__init__.py", line 49, in <module>
    from tensorflow.python import pywrap_tensorflow
  File "C:\Users\ranji\AppData\Local\Programs\Python\Python36\lib\site-packages\tensorflow\python\pywrap_tensorflow.py", line 30, in <module>
    self_check.preload_check()
  File "C:\Users\ranji\AppData\Local\Programs\Python\Python36\lib\site-packages\tensorflow\python\platform\self_check.py", line 82, in preload_check
    % (build_info.cudart_dll_name, build_info.cuda_version_number))
ImportError: Could not find 'cudart64_90.dll'. TensorFlow requires that this DLL be installed in a directory that is named in your %PATH% environment variable. Download and install CUDA 9.0 from this URL: https://developer.nvidia.com/cuda-toolkit

【问题讨论】：

标签： python windows tensorflow

【解决方案1】：

目前 Tensorflow 只支持 CUDA 9.0

但是您已经安装了 CUDA 9.1。您现在有两个选择：

选项 1：从源代码编译 Tensorflow，然后启用对 CUDA 9.1 版的支持
选项 2：您安装 CUDA 9.0 并等待 Tensorflow 支持 CUDA 9.1

我经历了同样的痛苦并决定安装 CUDA 9.0。这可能不那么麻烦。

有关 CUDA 9.1 的 Tensorflow 支持的当前状态以及如何从源代码编译 Tensorflow，另请参阅此 issue on GitHub。

【讨论】：