Recently in the study of data mining related knowledge, the class has mentioned keras related knowledge, under the class would like to build their own keras, helpless related information too little.
So he wrote this blog, for small white installation learning.
Keras is a deep learning framework based on Theano, designed to refer to torch, written in Python, is a highly modular neural network library that supports GPU and CPU, and is particularly simple to use for rapid development.
The entire Keras installation is more concise. The entire installation can refer to Keras Chinese documents:
http://keras-cn.readthedocs.io/en/latest/
The first is to install the development pack:
sudo apt install-y python-dev python-pip python-nose gcc g++ git gfortran vim
These are some of the basic tools that Keras needs.
After these downloads, you will need to continue downloading some of the operational acceleration Libraries:
sudo apt install-y libopenblas-dev liblapack-dev libatlas-base-dev
After all the above installation is complete, the next step is to install the Keras:
In the terminal, enter:
>>> sudo pip install-u--pre pip setuptools wheel
>>> sudo pip install-u--pre numpy scipy matplot Lib Scikit-learn scikit-image
>>> sudo pip install-u--pre theano
>>> sudo pip install-u--pre Keras
After installation, enter Python to test the success of the import related library.
>>> Import Theano
>>> import Keras
The installation process in official documents is cumbersome. There is also a way to install a more concise.
The first step is to install anaconda2 directly. This is equivalent to installing all the base class libraries in Keras.
Second, if you enter Pip install Keras in Anaconda2, you can install successfully.
Note: In the installation of ANACONDA2, the official website installation speed is really catching, can be downloaded through the mirror site, the speed will be much faster, the Web site is as follows:
https://mirrors.tuna.tsinghua.edu.cn/anaconda/archive/
The entire installation process can learn a lot of Linux operating system commands, for further study still have some help.