Installation of individual packages in Python, installing WHL files with PIP

Before the installation, has been more confused how to use, after installation, found that it is so simple

First I generally use is python27, official website download Python27msi installs WINDOW7 64 bit, already brought the PIP and Easy_install

But I don't know how to install the WHL directory,

After searching, it is not difficult to add Pip and Easy_install to the environment variable, as follows:

Copy scripts this directory directly to "right-click Computer-Properties-Advanced system settings-environment variables-System Variables-path-edit It"

Note that if there is no semicolon at the back, add a semicolon-in order to distinguish between different variables.

OK, now go to cmd, enter the PIP command is ready to use

Most of the packages in the Python learning process can be found under this connection, providing py27 or py3 WHL packages, connecting http://www.lfd.uci.edu/~gohlke/pythonlibs/, and the following pages

Just click on the package you want to download, and the very best thing is to tell you the package you want to use (require), which you need to download and install the auxiliary packages.

Matplotlib is a simple drawing often used, he needs numpy, Dateutil,pytz and other packages, download down to a certain path, attention is WHL format.

When I have installed the auxiliary package that I need before, install the Matplotlib package,

First jump to our download WHL folder path, directly with the PIP install file name can be (note with the tab-only need to knock the first few letters of the file name can-distinguish the other files-tab will automatically fill the file name) or all of their own beating, it is not the eyes are spent.

OK, that's enough, you can try drawing a simple diagram in the Wing IDE.

OK, now you can do what you want with Python's extra bag!

Recently want to learn TensorFlow, has been using Python, so want to borrow TF in Python expansion package.

The main problem is that TF only to python3.5 support better, other versions of Python is not to say no support, but the author is limited to the ability, many attempts have failed to succeed.

Below to get to the chase:

1. Download Anaconda, the author of the writing version is Anaconda4.3.0, the default comes with python3.6;

2. As previously stated, TensorFlow's current support for Python is only friendly to the 3.5 version, so we need to create a python3.5 environment with Anaconda.

Open cmd and enter:

[Plain]View PlainCopyprint?

1. Conda create-n tensorflow python=3.5 #环境名称为tensorflow The python version chosen is 3.5

Conda create-n tensorflow python=3.5  #环境名称为tensorflow The python version chosen is 3.5

3. Search Anaconda Navigator In the Start menu and click Run;

4. Click on the left side of the environments, you can see the "TensorFlow" Environment has been created;

5. Back to CMD and start the environment TensorFlow:

[Plain]View PlainCopy print?

1. Activate TensorFlow

Activate TensorFlow

6. Install the GPU version of TensorFlow:

[HTML]View PlainCopyprint?

1. Pip Install Tensorflow-gpu

Pip Install Tensorflow-gpu

To install the cup version of TensorFlow:

[Plain]View PlainCopy print?

1. Pip Install TensorFlow

Pip Install TensorFlow

7. Go back to Anaconda Navigator, enter the TensorFlow environment, install the Spyder plugin and enter the Spyder to

[Python]View PlainCopyprint?

1. Import TensorFlow as TF

Import TensorFlow as TF

#-----------------------------------------------------------

If there are problems with the directory not found in the installation, you can try to run cmd in Administrator mode and then perform the operation again.

Installation of individual packages in Python, installing WHL files with PIP