tensorflow的报错以及解决方法
安装tensorflow中遇见的问题
- Tensorflow-CPU
- pip Install tensorflow报错
- pip install tensorflow 之后出现的问题
- 测试是否安装成功
- tensorflow 简单的小例子、教程
Tensorflow-CPU
由于初次使用tensorflow，GPU版本需要安装cuda，cudnn等，太复杂，因此先安装cpu版的，最为最初的尝试，随着对tensorflow了解更多，后续可以尝试安装Tensorflow-GPU.
pip Install tensorflow报错
Fatal error in launcher: Unable to create process using '"c:……‘
此时需要强制升级pip，使用下列命令即可：
python -m pip install --upgrade --force-reinstall tensorflow
还可以升级tensorflow：
pip install --upgrade --force-reinstall tensorflow
查看tensorflow的版本：
tf.__version__
pip install tensorflow 之后出现的问题
pip成功之后，import tensorflow as tf 报错
Traceback (most recent call last):
File "C:\mysoft\Python\Python35\lib\site-packages\tensorflow\python\platform\self_check.py", line 47, in preload_check
ctypes.WinDLL(build_info.msvcp_dll_name)
File "C:\mysoft\Python\Python35\lib\ctypes\__init__.py", line 347, 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 "", line 1, in
import tensorflow
File "C:\mysoft\Python\Python35\lib\site-packages\tensorflow\__init__.py", line 24, in
from tensorflow.python import *
File "C:\mysoft\Python\Python35\lib\site-packages\tensorflow\python\__init__.py", line 49, in
from tensorflow.python import pywrap_tensorflow
File "C:\mysoft\Python\Python35\lib\site-packages\tensorflow\python\pywrap_tensorflow.py", line 30, in
self_check.preload_check()
File "C:\mysoft\Python\Python35\lib\site-packages\tensorflow\python\platform\self_check.py", line 55, in preload_check
% build_info.msvcp_dll_name)
ImportError: Could not find 'msvcp140.dll'. TensorFlow requires that this DLL be installed in a directory that is named in your %PATH% environment variable. You may install this DLL by downloading Visual C++ 2015 Redistributable Update 3 from this URL: https://www.microsoft.com/en-us/download/details.aspx?id=53587
其中，
ImportError: Could not find 'msvcp140.dll'. TensorFlow requires that this DLL be installed in a directory that is named in your %PATH% environment variable. You may install this DLL by downloading Visual C++ 2015 Redistributable Update 3 from this URL: https://www.microsoft.com/en-us/download/details.aspx?id=53587
原因：缺少VC++ 2015.
解决方法：安装vc++2015, URL：后面的即为下载vc++2015的地址,复制该链接，即可下载，大约14M，下载之后，安装即可,bingo!!
ps:（安装非常简单，1min不到，我之前因为担心又要安装大程序，迟迟不肯定下载并安装，尝试找其他方法解决，然后并没有找到其他的有效解决方法）
测试是否安装成功
经过上面的步骤，import tensorflow as tf已经不再报错，成功安装，因此进行测试：
import tensorflow as tf
hi = tf.constant('Hello, tensorflow')
sess = tf.Session()
这个时候，出现了个warning：
I T:\src\github\tensorflow\tensorflow\core\platform\cpu_feature_guard.cc:140] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2
查了资料，说这个问题warning不影响正常运行，可以忽略不计，但是还是想把这个问题解决。
解决方法：
- https://github.com/fo40225/tensorflow-windows-wheel 在这里，根据pyhton版本，找到对应的tensorflow的版本，下载下来,我是python3.6,对应的cpu的avx2的tesorflow版本下载下来的名字是tensorflow-1.12.0-cp36-cp36m-win_amd64.whl
- 把原来的版本的tensorflow卸载，pip uninstall tensorflow
- 安装新的tensorflow， pip install tensorflow-1.12.0-cp36-cp36m-win_amd64.whl，即安装上合适的tensorflow
- 重新测试，发现之前的warning已经消息
目前为止，tensorflow已经初步安装成功
可以开始运行tensorflow的一些简单的小程序
tensorflow 简单的小例子、教程
https://tensorflow.google.cn/tutorials/ 官网？
https://www.jianshu.com/p/2ea7a0632239 五分钟带你入门tensorflow
https://blog.csdn.net/geyunfei_/article/details/78782804 第一个入门demo
https://www.bilibili.com/video/av35974848/?spm_id_from=333.788.videocard.0 tensorflow的一个完整的中文教程，炼数成金的课程