Home PC Games Linux Windows Database Network Programming Server Mobile  
  Home \ Linux \ Installation of Theano + CUDA under Ubuntu     - How UTorrent download the ISO image from the command line (Linux)

- Node.js installed on Ubuntu Upstream version (Linux)

- Modify Linux SSH default port 22 in several ways (Linux)

- Acting on JavaScript events (Programming)

- Generated characters using Java Videos (Programming)

- Linux System Getting Started Learning: the local time in between Perl and UNIX timestamp conversion (Linux)

- The basic principles for the protection of a good linux server security (Linux)

- MySQL multi-master multi-cluster deployment uses alive Galera (Database)

- Detailed driver compiled into the Linux kernel (Programming)

- To install Git on Ubuntu systems (Server)

- Configuring DNS process under CentOS 6.5 (Server)

- Docker command Detailed (Linux)

- Change all child files and subdirectories in the owner permissions Ubuntu (Linux)

- Ubuntu cut screen method (Linux)

- Create Your Own Docker base image in two ways (Linux)

- Java 8 stream parsed into SQL (Programming)

- A key installation Gitlab 7 on RHEL6.4 and Setup Mail TX (Linux)

- Do not enter password login ssh (Server)

- MySQL flip-flop (Database)

- Linux, Chrome browser font settings beautification (Linux)

  Installation of Theano + CUDA under Ubuntu
  Add Date : 2017-08-31      
  Because of the need to use GPU accelerated learning machine learning algorithms, you need to install theano + cuda. Open source library for me is a big problem, that is difficult to install.

In order to improve this configuration, I spent three days back and forth, reloading three times Ubuntu reinstall the drivers five times to get it.

It made Citie, the most simple to install (if I remember it).

Because the next ubuntu various software relies too much relationship, it is best to install the new ubuntu theano and cuda, then update

The best is to do so, more likely to step wrong ..................

-------------------------------------------------- ------------------------------

First, install ubuntu

Under Windows installation theano and other dependent libraries rather complex, and even need to use edu mailbox.

Decisively abandoned, put up with ubuntu environment.

1, U disk production

After just great ubuntu (12.04) of iso, come up with a 4GU disk, insert formatted.

Open until the next step, the selected system, iso selected file.

Create after a while just fine.

2, to make room.

Systems need to free up some space to install ubuntu, preferably freed in more free hard disk.

I F disk to free up 20GB do Ubuntu system disk.


My computer - right click - Management - Storage - Disk Management

Right on the idle partition - compressed volume - Enter the vacated space size (MB) - Compression

3, restart the system

Boot into the boot from the selected boot disk U (press F2, or anything else, to see the boot motherboard interface), save and exit.

What then installed, it looked like a graphical interface.

ubuntu will automatically install to that vacated partition.

-------------------------------------------------- ------------------------------

Second, the installation Theano

1, installed after Ubuntu, let the computer access to the Internet.

(Campus network inode students do not toss, system settings - network -801.2x certification, set up a password and IP user can access)

2, after the Internet, open the software center, in the editing - source software, for the source of a good point (or 163 Education Network (if you are, then education network))

ctrl + alt + T to open the control terminal, enter: sudo apt-get update, so that the source can work.

3, enter the password, wait a minute, after a good turn input

1.sudo apt-get install python-numpy python-scipy python-dev python-pip python-nose g ++ libopenblas-dev git

2.sudo pip install Theano

And so on, well enough (if pip install a timeout, you can change the source or change the time period installation)

4, you can use

1.NumPy (~ 30s): python -c "import numpy; numpy.test ()"

2.SciPy (~ 1m): python -c "import scipy; scipy.test ()"

3.Theano (~ 30m): python -c "import theano; theano.test ()"

These three programs can run dynamic testing, no error on the line.

Third, the installation cuda

This is a tutorial to install the driver, but it cuda driver installation as I was successful.

The ninth step tutorial to do to restart the line.

Then add the environment variable :( Open with gedit .bashrc)

sudo gedit .bashrc

Add the following two lines

export PATH = / usr / local / cuda-5.5 / bin: $ PATH

export LD_LIBRARY_PATH = / usr / local / cuda-5.5 / lib64: $ LD_LIBRARY_PATH

Save, reboot.
- Linux Network Programming - signal blocking and shielding (block, unblock) (Programming)
- OpenSUSE / Linux network configuration (Linux)
- CentOS How quickly customize kernel binary RPM package (Linux)
- Nmcli based network management command-line tool (Linux)
- How to install and use the malware detection tool LMD and ClamAV antivirus engine on Linux (Linux)
- How to use jgit to manage Git submodule (Linux)
- Linux operating system security tools of the Logs (Linux)
- Linux installed Cisco Packet Tracer (Linux)
- Linux system security configuration (Linux)
- CentOS 6.5 upgrade to CentOS 7 (Linux)
- Linux use logs to troubleshoot (Linux)
- Linux, rename the file or folder (mv command and rename command) (Linux)
- Django 1.8 TEMPLATE_DIRS configuration and STATICFILES_DIRS (Server)
- Ubuntu uses under KVM + Qemu virtual machine build (Linux)
- 10 Linux in the passwd command examples (Linux)
- redis main building and disaster recovery from a cluster deployment (Database)
- Oracle database physical file backup / restore (Database)
- iptables using summary (Linux)
- imp / exp Oracle Database import and export commands (Database)
- Explore Android SQLite3 multithreading (Programming)
  CopyRight 2002-2022 newfreesoft.com, All Rights Reserved.