如何将虚拟环境设置为运行 tensorflow 基准测试

Question

我的目的是为运行 tensorflow benchmark 设置虚拟环境，但我在运行宁以下代码后的最后一条语句中遇到错误。请问我应该怎么做才能解决这个问题？仅供参考：我正在使用 Python 3.6.7，仅安装了 pip 和 virtualenv 包以及 CUDA 9.2/cuDNN7（系统路径中的所有 dll）。

virtualenv venv
venv\scripts\activate
#now you should be in virtual environment
#install using pip
pip install tf-nightly-gpu

python
import tensorflow

错误信息：

>>> import tensorflow
Traceback (most recent call last):
  File "C:\MyWork\testarea\testvirtualenv\venv\lib\site-packages\tensorflow\python\pywrap_tensorflow.py", line 58, in <module>
    from tensorflow.python.pywrap_tensorflow_internal import *
  File "C:\MyWork\testarea\testvirtualenv\venv\lib\site-packages\tensorflow\python\pywrap_tensorflow_internal.py", line 28, in <module>
    _pywrap_tensorflow_internal = swig_import_helper()
  File "C:\MyWork\testarea\testvirtualenv\venv\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:\MyWork\testarea\testvirtualenv\venv\lib\imp.py", line 243, in load_module
    return load_dynamic(name, filename, file)
  File "C:\MyWork\testarea\testvirtualenv\venv\lib\imp.py", line 343, 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:\MyWork\testarea\testvirtualenv\venv\lib\site-packages\tensorflow\__init__.py", line 24, in <module>
    from tensorflow.python import pywrap_tensorflow  # pylint: disable=unused-import
  File "C:\MyWork\testarea\testvirtualenv\venv\lib\site-packages\tensorflow\python\__init__.py", line 49, in <module>
    from tensorflow.python import pywrap_tensorflow
  File "C:\MyWork\testarea\testvirtualenv\venv\lib\site-packages\tensorflow\python\pywrap_tensorflow.py", line 74, in <module>
    raise ImportError(msg)
ImportError: Traceback (most recent call last):
  File "C:\MyWork\testarea\testvirtualenv\venv\lib\site-packages\tensorflow\python\pywrap_tensorflow.py", line 58, in <module>
    from tensorflow.python.pywrap_tensorflow_internal import *
  File "C:\MyWork\testarea\testvirtualenv\venv\lib\site-packages\tensorflow\python\pywrap_tensorflow_internal.py", line 28, in <module>
    _pywrap_tensorflow_internal = swig_import_helper()
  File "C:\MyWork\testarea\testvirtualenv\venv\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:\MyWork\testarea\testvirtualenv\venv\lib\imp.py", line 243, in load_module
    return load_dynamic(name, filename, file)
  File "C:\MyWork\testarea\testvirtualenv\venv\lib\imp.py", line 343, 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.
>

Answer 1

刚写完题就找到了答案。

在将 cudnn64_7.dll（CUDA 9.0 的 cuDNN 7.3.1）复制到 CUDA bin 目录之前将 CUDA 9.2 降级到 9.0 使其工作。

最后，以下脚本在两台不同的机器上产生了两个不同的结果。

virtualenv venv
venv\scripts\activate
pip install tf-nightly-gpu    
python tf_cnn_benchmarks.py --num_gpus=1 --batch_size=32 --model=resnet50 --variable_update=parameter_server

第 7 代四核 i7 与 GTX 970

总计 images/sec: 78.16

AMD 1920X 与 GTX 1080

总计 images/sec: 147.77

如何将虚拟环境设置为运行 tensorflow 基准测试

How to setup virtual environment to run tensorflow benchmarks

python

benchmarking

tensorflow

如何将虚拟环境设置为 运行 tensorflow 基准测试

How to setup virtual environment to run tensorflow benchmarks

python

benchmarking

tensorflow

如何将虚拟环境设置为运行 tensorflow 基准测试