Install NNI¶
NNI requires Python >= 3.7. It is tested and supported on Ubuntu >= 18.04, Windows 10 >= 21H2, and macOS >= 11.
There are 3 ways to install NNI:
Using pip¶
NNI provides official packages for x86-64 CPUs. They can be installed with pip:
pip install nni
Or to upgrade to latest version:
pip install --latest nni
You can check installation with:
nnictl --version
On Linux systems without Conda, you may encounter bash: nnictl: command not found
error.
In this case you need to add pip script directory to PATH
:
echo 'export PATH=${PATH}:${HOME}/.local/bin' >> ~/.bashrc
source ~/.bashrc
Installing from Source Code¶
NNI hosts source code on GitHub.
NNI has experimental support for ARM64 CPUs, including Apple M1. It requires to install from source code.
See Build from Source.
Using Docker¶
NNI provides official Docker image on Docker Hub.
docker pull msranni/nni
Installing Extra Dependencies¶
Some built-in algorithms of NNI requires extra packages.
Use nni[<algorithm-name>]
to install their dependencies.
For example, to install dependencies of DNGO tuner
:
pip install nni[DNGO]
This command will not reinstall NNI itself, even if it was installed in development mode.
Alternatively, you may install all extra dependencies at once:
pip install nni[all]
NOTE: SMAC tuner depends on swig3, which requires a manual downgrade on Ubuntu:
sudo apt install swig3.0
sudo rm /usr/bin/swig
sudo ln -s swig3.0 /usr/bin/swig