Get Started

Overview

There are multiple ways to get mlpack up and running. Python bindings can be installed using pip or conda, or built from source (see “Build from source” tutorials). For C++, if mlpack is not available via your preferred OS package manager, or if you need to build your own version (e.g. to apply optimizations, use a different set of BLAS/LAPACK, or build a different configuration), please refer to the “Build from source” tutorials. For Windows, prebuilt binaries will help you get started without the need of building mlpack. These packages include both the C++ mlpack library as well as the CLI tools.

Here is a summary of the currently available distribution options you can use depending on your needs:

Python

Get via pip or conda

pip install mlpack
conda install -c conda-forge mlpack

Ubuntu, Debian, Fedora, Red Hat

Install using a package manager

Build from source

Windows

Download prebuilt binaries

Install using a package manager

Build from source

macOS

Install using homebrew

brew install mlpack

Build from source