Deep Dream：CPU 工作但 GPU 不工作，无法加载原生 TensorFlow 运行时

Question

我对此很陌生，如果我的术语不完美，请道歉，但我已经做了很多尝试和搜索，但我无法通过这个问题。

我一直在尝试创建 Deep Dream 风格的传输图像，以便将它们批量处理为视频。

我尝试了一大堆不同的东西、指南和版本，我无法在 python 3.6.6 或更高版本上安装 tensorflow。 但遵循本指南：

(https://www.reddit.com/r/deepdream/comments/6z5tyz/guide_on_how_to_setup_neural_style_transfer_on/)

我终于让它工作了，但只是作为张量流 CPU。这很棒，但速度很慢。我有一个 GTX 1080，我想充分利用它。

当我安装 Tensorflow-gpu 时，它说它工作正常

C:\Users\kurti>pip3 install --upgrade tensorflow-gpu
Collecting tensorflow-gpu
  Using cached https://files.pythonhosted.org/packages/01/ac/ed64953d36b6516dc4a4ae983d7759bdcdaedd65541ee92a116909a524a8/tensorflow_gpu-1.13.1-cp35-cp35m-win_amd64.whl
Requirement already up-to-date: gast>=0.2.0 in c:\users\kurti\appdata\local\programs\python\python35\lib\site-packages (from tensorflow-gpu)
Requirement already up-to-date: keras-applications>=1.0.6 in c:\users\kurti\appdata\local\programs\python\python35\lib\site-packages (from tensorflow-gpu)
Requirement already up-to-date: keras-preprocessing>=1.0.5 in c:\users\kurti\appdata\local\programs\python\python35\lib\site-packages (from tensorflow-gpu)
Requirement already up-to-date: absl-py>=0.1.6 in c:\users\kurti\appdata\local\programs\python\python35\lib\site-packages (from tensorflow-gpu)
Requirement already up-to-date: protobuf>=3.6.1 in c:\users\kurti\appdata\local\programs\python\python35\lib\site-packages (from tensorflow-gpu)
Requirement already up-to-date: termcolor>=1.1.0 in c:\users\kurti\appdata\local\programs\python\python35\lib\site-packages (from tensorflow-gpu)
Requirement already up-to-date: wheel>=0.26 in c:\users\kurti\appdata\local\programs\python\python35\lib\site-packages (from tensorflow-gpu)
Requirement already up-to-date: astor>=0.6.0 in c:\users\kurti\appdata\local\programs\python\python35\lib\site-packages (from tensorflow-gpu)
Requirement already up-to-date: tensorflow-estimator<1.14.0rc0,>=1.13.0 in c:\users\kurti\appdata\local\programs\python\python35\lib\site-packages (from tensorflow-gpu)
Requirement already up-to-date: numpy>=1.13.3 in c:\users\kurti\appdata\local\programs\python\python35\lib\site-packages (from tensorflow-gpu)
Requirement already up-to-date: six>=1.10.0 in c:\users\kurti\appdata\local\programs\python\python35\lib\site-packages (from tensorflow-gpu)
Requirement already up-to-date: grpcio>=1.8.6 in c:\users\kurti\appdata\local\programs\python\python35\lib\site-packages (from tensorflow-gpu)
Requirement already up-to-date: tensorboard<1.14.0,>=1.13.0 in c:\users\kurti\appdata\local\programs\python\python35\lib\site-packages (from tensorflow-gpu)
Requirement already up-to-date: h5py in c:\users\kurti\appdata\local\programs\python\python35\lib\site-packages (from keras-applications>=1.0.6->tensorflow-gpu)
Requirement already up-to-date: setuptools in c:\users\kurti\appdata\local\programs\python\python35\lib\site-packages (from protobuf>=3.6.1->tensorflow-gpu)
Requirement already up-to-date: mock>=2.0.0 in c:\users\kurti\appdata\local\programs\python\python35\lib\site-packages (from tensorflow-estimator<1.14.0rc0,>=1.13.0->tensorflow-gpu)
Requirement already up-to-date: werkzeug>=0.11.15 in c:\users\kurti\appdata\local\programs\python\python35\lib\site-packages (from tensorboard<1.14.0,>=1.13.0->tensorflow-gpu)
Requirement already up-to-date: markdown>=2.6.8 in c:\users\kurti\appdata\local\programs\python\python35\lib\site-packages (from tensorboard<1.14.0,>=1.13.0->tensorflow-gpu)
Requirement already up-to-date: pbr>=0.11 in c:\users\kurti\appdata\local\programs\python\python35\lib\site-packages (from mock>=2.0.0->tensorflow-estimator<1.14.0rc0,>=1.13.0->tensorflow-gpu)
Installing collected packages: tensorflow-gpu
Successfully installed tensorflow-gpu-1.13.1

但是当我通过导入 GPU 版本进行测试时，我得到了这个：

C:\Users\kurti>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\kurti\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\pywrap_tensorflow.py", line 58, in <module>
    from tensorflow.python.pywrap_tensorflow_internal import *
  File "C:\Users\kurti\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\pywrap_tensorflow_internal.py", line 28, in <module>
    _pywrap_tensorflow_internal = swig_import_helper()
  File "C:\Users\kurti\AppData\Local\Programs\Python\Python35\lib\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 "C:\Users\kurti\AppData\Local\Programs\Python\Python35\lib\imp.py", line 242, in load_module
    return load_dynamic(name, filename, file)
  File "C:\Users\kurti\AppData\Local\Programs\Python\Python35\lib\imp.py", line 342, in load_dynamic
    return _load(spec)
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 "<stdin>", line 1, in <module>
  File "C:\Users\kurti\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\__init__.py", line 24, in <module>
    from tensorflow.python import pywrap_tensorflow  # pylint: disable=unused-import
  File "C:\Users\kurti\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\kurti\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\pywrap_tensorflow.py", line 74, in <module>
    raise ImportError(msg)
ImportError: Traceback (most recent call last):
  File "C:\Users\kurti\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\pywrap_tensorflow.py", line 58, in <module>
    from tensorflow.python.pywrap_tensorflow_internal import *
  File "C:\Users\kurti\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\pywrap_tensorflow_internal.py", line 28, in <module>
    _pywrap_tensorflow_internal = swig_import_helper()
  File "C:\Users\kurti\AppData\Local\Programs\Python\Python35\lib\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 "C:\Users\kurti\AppData\Local\Programs\Python\Python35\lib\imp.py", line 242, in load_module
    return load_dynamic(name, filename, file)
  File "C:\Users\kurti\AppData\Local\Programs\Python\Python35\lib\imp.py", line 342, in load_dynamic
    return _load(spec)
ImportError: DLL load failed: The specified module could not be found.


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.

CPU 工作的事实表明它与 CUDA、cuDNN 或 Tensor flow GPU 的版本有关，但我已经尝试过

Cuda 8(update2) + CuDNN 6
Cuda 8(update2) + CuDNN 5.1
Cuda 7.5 + cuDNN 5.1

之前还有一大堆较新的。 我相信我已经安装了所有正确的依赖项。

我已经检查了 PATH，就像这里有类似问题的其他人所建议的那样，看起来一切都很好，所以我不知道从这里去哪里？

我有这个用于 CPU 的工作，但我需要能够为视频做数百个这样的工作，所以非常感谢任何帮助。

---

更新：我发现这个不错的自检小脚本适用于 1.4 之前的版本 link

这表示它找不到我安装的 cuDNN v6，然后测试，仍然是同样的错误。再次运行自检，得到了这个结果

C:\Users\kurti\Documents\Projects SSD\TESTS\Deep Dream>tensorflow_self_check.py
ERROR: Failed to import the TensorFlow module.

WARNING! This script is no longer maintained!
=============================================

Since TensorFlow 1.4, the self-check has been integrated with TensorFlow itself,
and any missing DLLs will be reported when you execute the `import tensorflow`
statement. The error messages printed below refer to TensorFlow 1.3 and earlier,
and are inaccurate for later versions of TensorFlow.

- Python version is 3.5.

- TensorFlow is installed at: C:\Users\kurti\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow

- All required DLLs appear to be present. Please open an issue on the
  TensorFlow GitHub page: https://github.com/tensorflow/tensorflow/issues

更新 2：

经过更多研究，建议使用 Anaconda，因为这在 Windows 中显然不太容易出现上述问题。 这些是我正在采取的步骤

1. conda create -n [名称] python=3.5
2. 激活[名称]
3. conda 安装 tensorflow-gpu
4. pip3 安装 (numpy-1.16.2+mkl-cp35-cp35m-win_amd64.whl)
5. pip3 安装 (opencv_python-3.4.5-cp35-cp35m-win_amd64.whl)
6. pip3 安装 (scipy-1.2.1-cp35-cp35m-win_amd64.whl)

当我键入时，这不会给我任何错误

python
>>>import tensorflow
>>>import numpy

但是当我尝试运行脚本（适用于 cpu 版本）时，它再也找不到 numpy

    (tf2) C:\Users\kurti\GitProjects\DeepStyle\Deep Style\neural-style-tf-master>neural_style.py --content_img littleme.jpg --style_imgs starry-night.jpg --max_iterations 100 --verbose
Traceback (most recent call last):
  File "C:\Users\kurti\GitProjects\DeepStyle\Deep Style\neural-style-tf-master\neural_style.py", line 1, in <module>
    import tensorflow as tf
  File "C:\Users\kurti\Anaconda3\lib\site-packages\tensorflow\__init__.py", line 24, in <module>
    from tensorflow.python import pywrap_tensorflow  # pylint: disable=unused-import
  File "C:\Users\kurti\Anaconda3\lib\site-packages\tensorflow\python\__init__.py", line 47, in <module>
    import numpy as np
ModuleNotFoundError: No module named 'numpy'

运行 pip3 列表显示 numpy， 运行 pip list 显示 numpy， 运行 conda list 显示错误的 numpy，

我相信为所需的 .whl 运行 pip3 安装不会将它们安装到我的 conda 环境中，这是否正确，我该如何解决？

Answer 1

我不确定这是否有帮助，但 Tensorflow 安装文档建议 Windows GPU 版本仅适用于 cuDNN64_7.dll，所以我猜那是 cuDNN v6？我只提到这一点，因为错误看起来找不到所需的.dll，也许是 cuDNN 错误？文档中有一些关于如何设置路径等的额外信息，让我知道你是怎么做的。

https://www.tensorflow.org/install/gpu#windows_setup