win10+anaconda3+python3.6+opencv3.1.0

The recent use of Windows system more, just want to win10 under the deep learning this aspect of the research, then need to configure a good environment Balabala a lot of things. Silently remember a note, is the so-called "good memory is not as bad as written."

1. Installing Anaconda

This is a relatively simple operation, directly on the official website download can: https://www.anaconda.com/download/#windows, Select the Windows version of the Python3.6-version, according to their own computer is 32-bit or 64-bit installation can (I am 64-bit, so I do not understand the 32-bit, there is a conflict, please understand)

2. Installing OPENCV3

The previous step has completed the installation of Anaconda and Python3 and now requires the installation of OPENCV3.
Install the OPENCV3 installation package in Anaconda First, then launch the Anaconda virtual environment and install it with the PIP command.
OPENCV3 installation package location in Anaconda:
http://www.lfd.uci.edu/~gohlke/pythonlibs/#opencv
Select the corresponding file

To start the Anaconda virtual environment:
1. Open Anaconda Navigatro
2. Create a virtual environment:

3. Open the command line terminal:

4, in the terminal input: (D:\PROJECT\OPENCV_PYTHON-3.3.1-CP36-CP36M-WIN_AMD64.WHL for just download the OPENCV installation file path)
Pip Install D:\PROJECT\OPENCV_PYTHON-3.3.1-CP36-CP36M-WIN_AMD64.WHL

5. Install NumPy and matplotlib:
Pip Install NumPy
Pip Install Matplotlib


6. Installation test:

Test success

3 Installing TensorFlow

The

First declares that this is the installation under the CPU TensorFlow, after all, there is no money to buy the GPU (Love yourself), compared to the installation under the GPU, the CPU is more simple, do not need to install Cuda and CUDNN.  
before installing, you need to make sure that:  
1, the Python version is 3.5 64 bits and above. (TensorFlow supports Python 3.6 from 1.2, the previous official is not supported)  
2, your pip version >= 8.1. Use Pip-v to view the current PIP version and upgrade pip with python-m pip install-u pip.  
3, there is a stable network connection.  
Because Google's people have already put TensorFlow into a PIP installation package, it is now possible to install TensorFlow with a normal installation package, which is to go to the command line and execute the following simple statement:  
Pip Install–upgrade tensorflow 

4, installation success test:  
test code as follows:

import tensorflow as tfa = tf.random_normal((100, 100))b = tf.random_normal((100, 500))c = tf.matmul(a, b)sess = tf.InteractiveSession()sess.run(c)

• 1
• 2
• 3
• 4
• 5
• 6

win10+anaconda3+python3.6+opencv3.1.0