Installing MXNet on OS X (Mac)
mxnet.io/get_started/osx_setup.html
MXNet currently supports Python, R, Julia, and Scala. For users of Python on Mac, MXNet provides a set of Git Bash scripts that installs all of the required MXNet dependencies and the MXNet library.
Prepare Environment for GPU Installation
This section is optional. Skip to next section if you don’t plan to use GPUs. If you plan to build with GPU, you need to set up the environment for CUDA and cuDNN.
First, download and installCUDA 8 toolkit.
Once you have the CUDA Toolkit installed you will need to set up the required environment variables by adding the following to your ~/.bash_profile file:
exportCUDA_HOME=/usr/local/cudaexportDYLD_LIBRARY_PATH="$CUDA_HOME/lib:$DYLD_LIBRARY_PATH"exportPATH="$CUDA_HOME/bin:$PATH"
Reload ~/.bash_profile file and install dependencies:
. ~/.bash_profile? ? brew install coreutils? ? brew tap caskroom/cask
Then downloadcuDNN 5.
Unzip the file and change to the cudnn root directory. Move the header files and libraries to your local CUDA Toolkit folder:
$ sudo mv include/cudnn.h /Developer/NVIDIA/CUDA-8.0/include/? ? $ sudo mv lib/libcudnn* /Developer/NVIDIA/CUDA-8.0/lib? ? $ sudo ln -s /Developer/NVIDIA/CUDA-8.0/lib/libcudnn* /usr/local/cuda/lib/
Now we can start to build MXNet.
Clone the MXNet source code repository to your computer and run the installation script. In addition to installing MXNet, the script installsHomebrew,Numpy,LibBLAS,OpenCV,Graphviz,NumPyandJupyter.
It takes around 5 to 10 minutes to complete the installation.
# Clone mxnet repository. In terminal, run the commands WITHOUT "sudo"git clone https://github.com/dmlc/mxnet.git ~/mxnet --recursive# If building with GPU, add configurations to config.mk file:cd~/mxnet? ? cp make/config.mk .echo"USE_CUDA=1">>config.mkecho"USE_CUDA_PATH=/usr/local/cuda">>config.mkecho"USE_CUDNN=1">>config.mk# Install MXNet for Python with all required dependenciescd~/mxnet/setup-utils? ? bash install-mxnet-osx-python.sh
You can view the installation script we just used to install MXNet for Pythonhere.
Installing MXNet is a two-step process:
Build the shared library from the MXNet C++ source code.
Install the supported language-specific packages for MXNet.
Note:To change the compilation options for your build, edit themake/config.mkfile and submit a build request with themakecommand.
Install the dependencies, required for MXNet, with the following commands:
OpenBLAS and homebrew/science (for linear algebraic operations)
OpenCV (for computer vision operations)
# Paste this command in Mac terminal to install Homebrew/usr/bin/ruby -e"$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)"# Insert the Homebrew directory at the top of your PATH environment variableexportPATH=/usr/local/bin:/usr/local/sbin:$PATH
brew update? ? brew install pkg-config? ? brew install graphviz? ? brew install openblas? ? brew tap homebrew/science? ? brew install opencv# For getting pipbrew install python# For visualization of network graphspip install graphviz# Jupyter notebookpip install jupyter
After you have installed the dependencies, pull the MXNet source code from Git and build MXNet to produce an MXNet library calledlibmxnet.so.
The file calledosx.mkhas the configuration required for building MXNet on OS X. First copymake/osx.mkintoconfig.mk, which is used by themakecommand:
git clone --recursive https://github.com/dmlc/mxnet ~/mxnetcd~/mxnet? ? cp make/osx.mk ./config.mkecho"USE_BLAS = openblas">> ./config.mkecho"ADD_CFLAGS += -I/usr/local/opt/openblas/include">> ./config.mkecho"ADD_LDFLAGS += -L/usr/local/opt/openblas/lib">> ./config.mkecho"ADD_LDFLAGS += -L/usr/local/lib/graphviz/">> ./config.mk? ? make -j$(sysctl -n hw.ncpu)
If building withGPUsupport, add the following configuration to config.mk and build:
echo"USE_CUDA = 1">> ./config.mkecho"USE_CUDA_PATH = /usr/local/cuda">> ./config.mkecho"USE_CUDNN = 1">> ./config.mk? ? make
Note:To change build parameters, editconfig.mk.
We have installed MXNet core library. Next, we will install MXNet interface package for the programming language of your choice:
Install the MXNet Package for Python
Next, we install Python interface for MXNet. Assuming you are in~/mxnetdirectory, run below commands.
# Install MXNet Python packagecdpython? ? sudo python setup.py install
Check if MXNet is properly installed.
# You can change mx.cpu to mx.gpupython? ? >>> import mxnet as mx? ? >>>a=mx.nd.ones((2, 3), mx.cpu())>>> print((a * 2).asnumpy())[[2.? 2.? 2.][2.? 2.? 2.]]
If you don’t get an import error, then MXNet is ready for python.
Note: You can update mxnet for python by repeating this step after re-buildinglibmxnet.so.
Install the MXNet Package for R
You have 2 options:
Building MXNet with the Prebuilt Binary Package
Building MXNet from Source Code
Building MXNet with the Prebuilt Binary Package
For OS X (Mac) users, MXNet provides a prebuilt binary package for CPUs. The prebuilt package is updated weekly. You can install the package directly in the R console using the following commands:
install.packages("drat",repos="https://cran.rstudio.com")drat:::addRepo("dmlc")install.packages("mxnet")
Building MXNet from Source Code
Run the following commands to install the MXNet dependencies and build the MXNet R package.
Rscript-e"install.packages('devtools', repo = 'https://cran.rstudio.com')"
cdR-package? ? Rscript -e"library(devtools); library(methods); options(repos=c(CRAN='https://cran.rstudio.com')); install_deps(dependencies = TRUE)"cd..? ? make rpkg
Note:R-package is a folder in the MXNet source.
These commands create the MXNet R package as a tar.gz file that you can install as an R package. To install the R package, run the following command, use your MXNet version number:
R CMD INSTALL mxnet_current_r.tar.gz
Install the MXNet Package for Julia
The MXNet package for Julia is hosted in a separate repository, MXNet.jl, which is available onGitHub. To use Julia binding it with an existing libmxnet installation, set theMXNET_HOMEenvironment variable by running the following command:
exportMXNET_HOME=//libmxnet
The path to the existing libmxnet installation should be the root directory of libmxnet. In other words, you should be able to find thelibmxnet.sofile at$MXNET_HOME/lib. For example, if the root directory of libmxnet is~, you would run the following command:
exportMXNET_HOME=/~/libmxnet
You might want to add this command to your~/.bashrcfile. If you do, you can install the Julia package in the Julia console using the following command:
Pkg.add("MXNet")
For more details about installing and using MXNet with Julia, see theMXNet Julia documentation.
Install the MXNet Package for Scala
Before you build MXNet for Scala from source code, you must completebuilding the shared library. After you build the shared library, run the following command from the MXNet source root directory to build the MXNet Scala package:
make scalapkg
This command creates the JAR files for the assembly, core, and example modules. It also creates the native library in thenative/{your-architecture}/targetdirectory, which you can use to cooperate with the core module.
To install the MXNet Scala package into your local Maven repository, run the following command from the MXNet source root directory:
make scalainstall
