The day Chainer runs on GPU CentOS edition (up to CPU)

be moved. I think I worked hard for about a week. In conclusion, I will write the important thing first.

-Update the Nvidia driver to the latest version. ・ First, put all the python libraries.

I would like to elaborate below.

What went

  1. Introduction of python 2.7.4   problem Since python 2.6.6 was installed, it was necessary to install 2.7.4.

** Solution **

$ su 
$ yum install zlib zlib-devel tk-devel tcl-devel sqlite-devel ncurses-devel gdbm-devel readline-devel bzip2-devel db4-devel openssl-devel
$ wget http://python.org/ftp/python/2.7.2/Python-2.7.4.tgz
$ tar zxvf Python-2.7.4.tgz 
$ cd Python-2.7.4
$ ./configure --with-threads --enable-shared
$ make
$ make install
$ ln -s /usr/local/lib/libpython2.7.so.1.0 /lib64/

** A word ** Make a strong symbolic link Do make install I put in various things, but I often do not know

** Supplement ** It seems that you can install it without changing the existing version by doing make alt install

** Referenced ** http://qiita.com/fuzzy31u/items/c35f2f4db5f3a48d86a7 http://a-records.info/2014/10/25/34/

  1. Introduction of various python libraries

problem When I installed Chainer, I found that various libraries were missing, so I installed it. I didn't include pip, so I included it.

** Solution **

$ su
$ yum install python-setuptools
$ wget http://peak.telecommunity.com/dist/ez_setup.py
$ /usr/local/bin/python ez_setup.py
$ easy_install pip
$ easy_install numpy
$ easy_install six
$ easy_install Mako
$ easy_install scipy
$ pip install scikit-learn

** A word ** When installing easy_install, install it from ez_setup.py Since / usr / bin / python is centos original python Later, you'll be worried about the python version

You can use easy_install or pip install Only scikit-learn was entered with pip install

** Referenced ** http://www.tomoyan.net/dokuwiki/python/easy_install http://inoccu.net/blog/2013/04/16/075618.html

  1. Introduction of chainer

problem Insert chainer

** Solution **

$ su
$ pip install chainer
$ git clone  https://github.com/pfnet/chainer
$ cd chainer/examples/mnist/
$ python train_mnist.py

** A word ** git clone https://github.com/pfnet/chainer Contains sample files. Now you can run it on the cpu version. From now on, I would like to explain learning using GPU.

Recommended Posts

The day Chainer runs on GPU CentOS edition (up to CPU)
I want to restart CentOS 8 on time every day.
Install Chainer on CentOS 6.7
Setting up grub on CentOS 8
Until docker-compose up on CentOS7
Install Chainer 1.6 (GPU) on Windows 7.
Settings to be made when starting up the linux server centos7
Script for backing up folders on the server to Google Drive
The easiest way to get Chainer v1.5 + CUDA + cuDNN on Windows
Mount the volume attached to CentOS 7
Steps to deploy EMLauncher on CentOS 8
Steps to install MySQL 8.0 on CentOS 8.1
Steps to install VirtualBox on CentOS
[Day3] Preparing to connect to the database
How to install PyPy on CentOS
Set up Python environment on CentOS
How to install TensorFlow on CentOS 7
How to install Maven on CentOS