在Pycharm中安装Tensorflow-gpu导致很多出错，
然后重新安装CUDA+CUDNN把环境搞崩了，
所以，整个环境重新安装——把Anaconda卸载了，Python卸载了，Pycharm卸载了。
Anaconda版本：Anaconda3-5.0.1-Windows-x86_64.exe
Pycharm版本：pycharm-community-2018.1.1.exe
CUDA版本：cuda_9.1.85_win10_network.exe
CuDNN版本：cudnn-9.1-windows10-x64-v7.1.zip
可参考https://blog.csdn.net/Eppley/article/details/79299075
conda install keras
E:\zql\data_processing_v1.0\pre-processed_data(unit_area)_Pieces>conda install keras
Fetching package metadata ...............
Solving package specifications: .
Package plan for installation in environment C:\Users\yzzn\Anaconda3:
The following NEW packages will be INSTALLED:
backports.weakref: 1.0rc1-py36_0 https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
keras: 2.1.5-py36_0 defaults
libprotobuf: 3.2.0-vc14_0 https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
markdown: 2.6.9-py36_0 https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
protobuf: 3.2.0-py36_0 https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
tensorflow: 1.2.1-py36_0 https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
The following packages will be UPDATED:
anaconda: 5.0.1-py36h8316230_2 defaults --> custom-py36_0 https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
vs2015_runtime: 14.0.25123-hd4c4e62_2 defaults --> 14.0.25420-0 https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
The following packages will be SUPERSEDED by a higher-priority channel:
bleach: 2.0.0-py36h0a7e3d6_0 defaults --> 1.5.0-py36_0 https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
conda: 4.3.30-py36h7e176b0_0 defaults --> 4.3.30-py36h7e176b0_0 https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
conda-env: 2.6.0-h36134e3_1 defaults --> 2.6.0-0 https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
html5lib: 0.999999999-py36ha09b1f3_0 defaults --> 0.9999999-py36_0 https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
Proceed ([y]/n)? y
conda-env-2.6. 100% |###############################| Time: 0:00:00 0.00 B/s
vs2015_runtime 100% |###############################| Time: 0:00:00 3.58 MB/s
anaconda-custo 100% |###############################| Time: 0:00:00 881.55 kB/s
libprotobuf-3. 100% |###############################| Time: 0:00:00 11.06 MB/s
markdown-2.6.9 100% |###############################| Time: 0:00:00 0.00 B/s
backports.weak 100% |###############################| Time: 0:00:00 0.00 B/s
html5lib-0.999 100% |###############################| Time: 0:00:00 11.66 MB/s
protobuf-3.2.0 100% |###############################| Time: 0:00:00 10.03 MB/s
bleach-1.5.0-p 100% |###############################| Time: 0:00:00 0.00 B/s
tensorflow-1.2 100% |###############################| Time: 0:00:02 10.59 MB/s
keras-2.1.5-py 100% |###############################| Time: 0:00:17 28.90 kB/s
conda-4.3.30-p 100% |###############################| Time: 0:00:00 11.11 MB/s
DEBUG menuinst_win32:__init__(185): Menu: name: 'Anaconda${PY_VER} ${PLATFORM}', prefix: 'C:\Users\yzzn\Anaconda3', env_name: 'None', mode: 'None', used_mode: 'user'
DEBUG menuinst_win32:create(299): Shortcut cmd is C:\Users\yzzn\Anaconda3\pythonw.exe, args are ['-m', 'webbrowser', '-t', '"https://anaconda.org/"']
E:\zql\data_processing_v1.0\pre-processed_data(unit_area)_Pieces>python
Python 3.6.3 |Anaconda custom (64-bit)| (default, Oct 15 2017, 03:27:45) [MSC v.1900 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import keras
Using TensorFlow backend.
>>>
conda install pillow
E:\zql\data_processing_v1.0\pre-processed_data(unit_area)_Pieces>conda install pillow
Fetching package metadata ...............
Solving package specifications: .
Package plan for installation in environment C:\Users\yzzn\Anaconda3:
The following packages will be SUPERSEDED by a higher-priority channel:
freetype: 2.8-vc14h17c9bdf_0 defaults --> 2.5.5-vc14_2 https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
matplotlib: 2.1.0-py36h11b4b9c_0 defaults --> 2.0.2-np113py36_0 https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
pillow: 4.2.1-py36hdb25ab2_0 defaults --> 4.2.1-py36_0 https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
tk: 8.6.7-vc14hb68737d_1 defaults --> 8.5.18-vc14_0 https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
Proceed ([y]/n)? y
tk-8.5.18-vc14 100% |###############################| Time: 0:00:00 5.40 MB/s
freetype-2.5.5 100% |###############################| Time: 0:00:00 10.27 MB/s
pillow-4.2.1-p 100% |###############################| Time: 0:00:00 7.95 MB/s
matplotlib-2.0 100% |###############################| Time: 0:00:00 10.46 MB/s
在Pycharm中安装keras库