Prerequisites

This package supports Linux, Mac OSX, and Windows platforms. You may also want to check:

• mxnet.
• mxnet-cu112 with CUDA-11.2 support.
• mxnet-cu110 with CUDA-11.0 support.
• mxnet-cu102 with CUDA-10.2 support.
• mxnet-cu101 with CUDA-10.1 support.
• mxnet-cu100 with CUDA-10.0 support.
• mxnet-native CPU variant without MKLDNN.

To use this package on Linux you need the libquadmath.so.0 shared library. On Debian based systems, including Ubuntu, run sudo apt install libquadmath0 to install the shared library. On RHEL based systems, including CentOS, run sudo yum install libquadmath to install the shared library. As libquadmath.so.0 is a GPL library and MXNet part of the Apache Software Foundation, MXNet must not redistribute libquadmath.so.0 as part of the Pypi package and users must manually install it.

Please also note that the AArch64 wheels are built with MKLDNN, and MKLDNN in turn is integrated with Arm Performance Libraries (APL). Because APL's license is also not compatible with Apache, users must manually install APL in their system.

To install for other platforms (e.g. Windows, Raspberry Pi/ARM) or other versions, check Installing MXNet for instructions on building from source.

Installation

To install, use:

pip install mxnet