pip（easy_install）anaconda本地安装文件包（附gluon-cv的升级过程）
pip本地安装
pip下载到本地后切换到下载目录再安装，小事，但是这样的包多了，很繁琐！
pip install piptestfile-1.1.0b0-py2.py3-none-any.whl（.whl安装包名称）
这里有个小窍门：只输入前面一部分，例如输入上图文件名中的开头字母T，然后按Tab键，后面部分会自动补全
anaconda本地安装
从某个地址安装
conda install -c https://conda.website.org/<package-name> <package-name>
本地安装
conda install --offline C:\abc-1.2.3-py35_0.tar.bz2
conda install --use-local c:/downloads/abc-1.2.3-py35_0.tar.bz2
貌似说conda本地安装不会添加依赖（参考：https://github.com/conda/conda/issues/1884）
附录： easy_install 和 pip的介绍
easy_install和pip都是用来下载安装Python一个公共资源库PyPI的相关资源包的，pip是easy_install的改进版，提供更好的提示信
息，删除package等功能。老版本的python中只有easy_install，没有pip。
easy_install 打包和发布 Python 包
pip 是包管理
easy_install 的安装
前提是python的环境已配置好，下载地址：https://pypi.python.org/pypi/ez_setup
下载 ez_setup.py 后，在命令 界面输入：python ez_setup.py
pip 的安装
待根据上述操作，安装好easy_install 之后，再安装pip
下载地址：https://pypi.python.org/pypi/pip
下载到桌面，解压，然后输入命令：python setup.py install ，开始安装。
easy_install的用法
这个工具我用的比较少，不如PIP那么频繁，举个例子，我把protobuf的egg下载到本地后，这样安装
python -m easy_install protobuf-3.6.1-py3.6-win-amd64.egg
pthon -m 的意思是run module as script (参考：python --help)
大体上的用法如下，
安装一个包
easy_install 包名
easy_install "包名 == 包的版本号"
升级一个包
easy_install -U "包名 >= 包的版本号"
pip 的用法
安装一个包
pip install 包名
pip install 包名 == 包的版本号
升级一个包 (如果不提供version号，升级到最新版本）
pip install --upgrade 包名 >= 包的版本号
删除一个包
pip uninstall 包名
强迫升降级安装
pip install 'package==1.0.0' --force-reinstall
配置pip3
我经常碰到找不到pip3的情况，一般，快速的解决办法是，把pip3.exe, easy_install-3.6.exe等需要用到文件直接拷贝到所在的环境目录，比如我的某个环境目录下是这样的，
D:\Anaconda3\envs\env_name_py36\Scripts
________
|---pip.exe
________
|---pip3.exe
________
|---easy_install.exe
________
|---easy_install-3.exe
也不知道总是这样直接拷贝有没有问题，总之缺什么的话也只能另外再补上安装了。
后记：
正常情况下，如果你能使用官网所提供的指令（https://gluon-cv.mxnet.io/install/install-more.html）一次安装成功，
pip install --upgrade gluoncvpip会帮你把相关的依赖统统都装好，加起来主要包括：
Installing collected packages: tqdm, portalocker, Pillow, cycler, kiwisolver, matplotlib, scipy, gluoncv
注意，官网上原有的指令是
pip install --upgrade mxnet gluoncv而实际上我不推荐和mxnet一起安装，单独安装mxnet你可以选择自己想要的版本，尤其是有gpu的情况下，版本要和驱动程序的版本一致，否则会产生相当多的问题。
情况不好的时候，就和我下面的方法一样，挨个自己装吧！
这主要是因为最近网络抽风严重，一个几十K的东西，费老半天装不上去，只好下载后再pip, 下面这个例子是升级gluon-cv的全过程，这时直接使用pip install --upgrade gluoncv 就别说了，根本装不上去，所以只好到这里（https://pypi.org/project/gluoncv/#files）把文件包下载下来，具体下载地址是：https://files.pythonhosted.org/packages/fa/81/37a00609cb53da3671adb106b9bc03fb1c029ad5a8db4bc668283e65703d/gluoncv-0.7.0-py2.py3-none-any.whl
然后再切换到下载目录下安装，如下
(mxnet36) H:\>cd H:\down_pipwheels
(mxnet36) H:\down_pipwheels>pip install gluoncv-0.7.0-py2.py3-none-any.whl --upgrade
Processing h:\down_pipwheels\gluoncv-0.7.0-py2.py3-none-any.whl
Requirement already satisfied, skipping upgrade: matplotlib in d:\anaconda3\envs\mx36\lib\site-packages (from gluoncv==0.7.0) (3.1.1)
Collecting portalocker (from gluoncv==0.7.0)
Downloading https://files.pythonhosted.org/packages/53/84/7b3146ec6378d28abc73ab484f09f47dfa008ad6f03f33d90a369f880e25/portalocker-1.7.0-py2.py3-none-any.whl
Requirement already satisfied, skipping upgrade: numpy in d:\anaconda3\envs\mx36\lib\site-packages (from gluoncv==0.7.0) (1.16.5)
Requirement already satisfied, skipping upgrade: tqdm in d:\anaconda3\envs\mx36\lib\site-packages (from gluoncv==0.7.0) (4.36.1)
Requirement already satisfied, skipping upgrade: Pillow in d:\anaconda3\envs\mx36\lib\site-packages (from gluoncv==0.7.0) (6.2.0)
Requirement already satisfied, skipping upgrade: scipy in d:\anaconda3\envs\mx36\lib\site-packages (from gluoncv==0.7.0) (1.3.1)
Requirement already satisfied, skipping upgrade: requests in d:\anaconda3\envs\mx36\lib\site-packages (from gluoncv==0.7.0) (2.22.0)
Requirement already satisfied, skipping upgrade: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in d:\anaconda3\envs\mx36\lib\site-packages (from matplotlib->gluoncv==0.7.0) (2.4.5)
Requirement already satisfied, skipping upgrade: python-dateutil>=2.1 in d:\anaconda3\envs\mx36\lib\site-packages (from matplotlib->gluoncv==0.7.0) (2.8.1)
Requirement already satisfied, skipping upgrade: cycler>=0.10 in d:\anaconda3\envs\mx36\lib\site-packages (from matplotlib->gluoncv==0.7.0) (0.10.0)
Requirement already satisfied, skipping upgrade: kiwisolver>=1.0.1 in d:\anaconda3\envs\mx36\lib\site-packages (from matplotlib->gluoncv==0.7.0) (1.1.0)
Requirement already satisfied, skipping upgrade: pywin32!=226; platform_system == "Windows" in d:\anaconda3\envs\mx36\lib\site-packages (from portalocker->gluoncv==0.7.0) (223)
Requirement already satisfied, skipping upgrade: chardet<3.1.0,>=3.0.2 in d:\anaconda3\envs\mx36\lib\site-packages (from requests->gluoncv==0.7.0) (3.0.4)
Requirement already satisfied, skipping upgrade: idna<2.9,>=2.5 in d:\anaconda3\envs\mx36\lib\site-packages (from requests->gluoncv==0.7.0) (2.8)
Requirement already satisfied, skipping upgrade: certifi>=2017.4.17 in d:\anaconda3\envs\mx36\lib\site-packages (from requests->gluoncv==0.7.0) (2020.4.5.1)
Requirement already satisfied, skipping upgrade: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in d:\anaconda3\envs\mx36\lib\site-packages (from requests->gluoncv==0.7.0) (1.24.2)
Requirement already satisfied, skipping upgrade: six>=1.5 in d:\anaconda3\envs\mx36\lib\site-packages (from python-dateutil>=2.1->matplotlib->gluoncv==0.7.0) (1.12.0)
Requirement already satisfied, skipping upgrade: setuptools in d:\anaconda3\envs\mx36\lib\site-packages (from kiwisolver>=1.0.1->matplotlib->gluoncv==0.7.0) (41.4.0)
Installing collected packages: portalocker, gluoncv
Found existing installation: gluoncv 0.5.0
Uninstalling gluoncv-0.5.0:
Successfully uninstalled gluoncv-0.5.0
Successfully installed gluoncv-0.7.0 portalocker-1.7.0
