Installing MXNet from source on Ubuntu

The following installation instructions are for building MXNet from source on computers running Ubuntu 16.04 or higher. For instructions to build MXNet from source on other platforms, see the general Build From Source guide.

Instead of building from source, you can install a binary version of MXNet. For that, please follow the information at Get Started.

Building MXNet from source is a two-step process:

  1. Build the shared library from the MXNet C++ source code.
  2. (optional) Install the supported language-specific packages for MXNet.


CUDA Dependencies

If you plan to build with GPU, you need to set up the environment for CUDA and cuDNN.

First, download and install CUDA toolkit. CUDA 9.2 is recommended.

Then download cuDNN 7.1.4.

Unzip the file and change to the cuDNN root directory. Move the header and libraries to your local CUDA Toolkit folder:

tar xvzf cudnn-9.2-linux-x64-v7.1
sudo cp -P cuda/include/cudnn.h /usr/local/cuda/include
sudo cp -P cuda/lib64/libcudnn* /usr/local/cuda/lib64
sudo chmod a+r /usr/local/cuda/include/cudnn.h /usr/local/cuda/lib64/libcudnn*
sudo ldconfig

Build the MXNet shared library from source

It is recommended that you review the general build from source instructions before continuing.

On Ubuntu versions 16.04 or later, you need the following dependencies:

Step 1: Install prerequisite packages.

sudo apt-get update
sudo apt-get install -y build-essential git ninja-build ccache libopenblas-dev libopencv-dev cmake

Instead of libopenblas-dev you may also choose a different math library. Further information is provided in the source guide's Math Library Selection section.

libopencv-dev is an optional dependency. You can delete it from above apt-get install line and build MXNet without OpenCV support by setting USE_OPENCV to OFF in the configuration file described below.

Note: CMake 3.13 or higher is required. If you are running an older version of CMake, you will see an error message like CMake 3.13 or higher is required. You are running version 3.10.2 during Step 3. In that case please update CMake on your system. You can download and install latest CMake from or via pip3 with pip3 install --user --upgrade "cmake>=3.13.2". If pip3 is not available on your system, you can install it via sudo apt-get install -y python3-pip. After installing cmake with pip3, it is usually available at ~/.local/bin/cmake or directly as cmake.

Step 2: Download MXNet sources and configure

Clone the repository:

git clone --recursive mxnet
cd mxnet
cp config/linux.cmake config.cmake  # or config/linux_gpu.cmake for build with CUDA

Please edit the config.cmake file based on your needs. The file contains a series of set(name value CACHE TYPE "Description") entries. You can change the values of the respective value.

For a GPU-enabled build make sure you have installed the CUDA dependencies first). When building a GPU-enabled build on a machine without GPU, MXNet build can't autodetect your GPU architecture and will target all available GPU architectures. Please set the MXNET_CUDA_ARCH variable in config.cmake to your desired cuda architecture to speed up the build.

To (optionally) build with MKL math library, please install MKL first based on the guide in Math Library Selection.

Step 3: Build MXNet core shared library.

rm -rf build
mkdir -p build && cd build
cmake ..
cmake --build .

Specify cmake --build . --parallel N to set the number of parallel compilation jobs. Default is derived from CPUs available.

After a successful build, you will find the in the build folder in your MXNet project root. is required to install language bindings described in the next section.

Installing Language Packages for MXNet

After you have installed the MXNet core library. You may install MXNet interface packages for the programming language of your choice: - Python - C++ - Clojure - Julia - Perl - R - Scala - Java

Install MXNet for Python

To install the MXNet Python binding navigate to the root of the MXNet folder then run the following:

cd python
pip install --user -e .

Note that the -e flag is optional. It is equivalent to --editable and means that if you edit the source files, these changes will be reflected in the package installed.

Optional Python Packages

You may optionally install graphviz library that is used for visualizing network graphs you build on MXNet. You may also install Jupyter Notebook which is used for running MXNet tutorials and examples.

pip install --user graphviz==0.8.4 jupyter

Install the MXNet Package for C++

Refer to the C++ Package setup guide.

Install the MXNet Package for Clojure

Refer to the Clojure setup guide.

Install the MXNet Package for Julia

Install Julia

The package available through apt-get is old and not compatible with the latest version of MXNet. Fetch the latest version (1.0.3 at the time of this writing).

wget -qO julia-10.tar.gz

Place the extracted files somewhere like a julia folder in your home dir.

mkdir ~/julia
mv julia-10.tar.gz ~/julia
cd ~/julia
tar xvf julia-10.tar.gz

Test Julia.

cd julia-1.0.3/bin
julia -e 'using InteractiveUtils; versioninfo()'

If you're still getting the old version, remove it.

sudo apt remove julia

Update your PATH to have Julia's new location. Add this to your .zshrc, .bashrc, .profile or .bash_profile.

export PATH=~/julia/julia-1.0.3/bin:$PATH

Validate your PATH.

echo $PATH

Validate Julia works and is the expected version.

julia -e 'using InteractiveUtils; versioninfo()'

Setup Your MXNet-Julia Environment

For each of the following environment variables, add the commands to your .zshrc, .bashrc, .profile or .bash_profile to make them persist.

Create a julia-depot folder and environment variable.

mkdir julia-depot
export JULIA_DEPOT_PATH=$HOME/julia/julia-depot

To use the Julia binding with an existing libmxnet installation, set the MXNET_HOME environment variable to the MXNet source root. For example:

export MXNET_HOME=$HOME/mxnet

Now set the LD_LIBRARY_PATH environment variable to where is found. If you can't find it, you might have skipped the building MXNet step. Go back and build MXNet first. For example:


Verify the location of and set the LD_PRELOAD environment variable. bash export LD_PRELOAD=/usr/lib/x86_64-linux-gnu/$LD_PRELOAD

With all of these updates, here's an example of what you might want to have in your .zshrc, .bashrc, .profile or .bash_profile.

export PATH=$HOME/bin:$HOME/.local/bin:$HOME/julia/julia-1.0.3/bin:$PATH
export JULIA_DEPOT_PATH=$HOME/julia/julia-depot
export MXNET_HOME=$HOME/mxnet
export LD_PRELOAD=/usr/lib/x86_64-linux-gnu/$LD_PRELOAD

Install MXNet with Julia:

julia --color=yes --project=./ -e \
      'using Pkg; \
       Pkg.develop(PackageSpec(name="MXNet", path = joinpath(ENV["MXNET_HOME"], "julia")))'

For more details about installing and using MXNet with Julia, see the MXNet Julia documentation.

Install the MXNet Package for Perl

Before you build MXNet for Perl from source code, you must complete building the shared library. After you build the shared library, run the following command from the MXNet source root directory to build the MXNet Perl package:

sudo apt-get install libmouse-perl pdl cpanminus swig libgraphviz-perl
cpanm -q -L "${HOME}/perl5" Function::Parameters Hash::Ordered PDL::CCS

export PERL5LIB=${HOME}/perl5/lib/perl5

cd ${MXNET_HOME}/perl-package/AI-MXNetCAPI/
perl Makefile.PL INSTALL_BASE=${HOME}/perl5
make install

cd ${MXNET_HOME}/perl-package/AI-NNVMCAPI/
perl Makefile.PL INSTALL_BASE=${HOME}/perl5
make install

cd ${MXNET_HOME}/perl-package/AI-MXNet/
perl Makefile.PL INSTALL_BASE=${HOME}/perl5
make install

Install the MXNet Package for R

Before you build MXNet for R from source code, you must complete building the shared library.

Minimum Requirements 1. GCC 4.8 or later to compile C++ 11. 2. GNU Make

Build the MXNet core shared library

Step 1 Install R, cran-devtools and dependencies

To install R and the devtools, run

sudo apt-get update
sudo apt-get install -y r-base-core r-cran-devtools libcairo2-dev libxml2-dev

libxml2-dev is required for the roxygen2 dependency and libcairo2-dev is required for the suggested imager dependency.

To generate documentation, it is also required to install roxygen2.

> install.packages("roxygen2")
> Would you like to use a personal library instead?  (y/n) y
> Would you like to create a personal library ... to install packages into?  (y/n) y

Note: To successfully complete the next step, you need a personal R library. If you were able to run install.packages("roxygen2") above, you either had already, or you have successfully created a personal library just now.

Step 2 Make and install the MXNet-R bindings.

make -f R-package/Makefile rpkg

Verify MXNet-R Installation

You can verify your MXNet-R installation as follows:


At the R prompt enter the following:

a <- mx.nd.ones(c(2,3), ctx = mx.cpu())
b <- a * 2 + 1

You should see the following output:

     [,1] [,2] [,3]
[1,]    3    3    3
[2,]    3    3    3
> quit()

Install the MXNet Package for Scala

To use the MXNet-Scala package, you can acquire the Maven package as a dependency.

Further information is in the MXNet-Scala Setup Instructions.

If you use IntelliJ or a similar IDE, you may want to follow the MXNet-Scala on IntelliJ tutorial instead.

Install the MXNet Package for Java

To use the MXNet-Java package, you can acquire the Maven package as a dependency.

Further information is in the MXNet-Java Setup Instructions.

If you use IntelliJ or a similar IDE, you may want to follow the MXNet-Java on IntelliJ tutorial instead.


You are more than welcome to contribute easy installation scripts for other operating systems and programming languages. See the community contributions page for further information.

Next Steps