To install EvalML on your platform, run one of the following commands:
$ pip install evalml
$ conda install -c conda-forge evalml
# See the EvalML with core dependencies only section $ pip install evalml --no-dependencies $ pip install -r core-requirements.txt
# See the EvalML with core dependencies only section $ conda install -c conda-forge evalml-core
EvalML with core dependencies only#
EvalML includes several optional dependencies. The
catboost packages support pipelines built around those modeling libraries. The
ipywidgets packages support plotting functionality in automl searches. These dependencies are recommended, and are included with EvalML by default but are not required in order to install and use EvalML.
EvalML’s core dependencies are listed in
core-requirements.txt in the source code, while the default collection of requirements is specified in
$ pip install evalml --no-dependencies $ pip install -r core-requirements.txt
$ conda install -c conda-forge evalml-core
EvalML allows users to install add-ons individually or all at once:
Update Checker: Receive automatic notifications of new EvalML releases
Time Series: Use EvalML with Facebook’s Prophet library for time series support.
$ pip install evalml[complete]
$ pip install evalml[prophet]
$ pip install evalml[update_checker]
$ conda install -c conda-forge alteryx-open-src-update-checker
Time Series support with Facebook’s Prophet#
To support the
Prophet time series estimator, be sure to install it as an extra requirement. Please note that this may take a few minutes.
Prophet is currently only supported via pip installation in EvalML for Mac with CmdStan as a backend.
pip install evalml[prophet]
Another option for installing Prophet with CmdStan as a backend is to use
Note: In order to do this, you must have the EvalML repo cloned and you must be in the top level folder
<your_directory>/evalml/ to execute this command.
This command will do the following:
install_cmdstan.pyscript found within your
STAN_BACKENDenvironment variables set.
site-packages path is incorrect or you’d like to specify a different one, just run
make installdeps-prophet SITE_PACKAGES_DIR="<path_to_your_site_packages>".
If you’d like to have more fine-tuned control over the installation steps for Prophet, such as specifying the backend, follow these steps:
$ pip install prophet==1.0.1
$ pip install cmdstanpy==0.9.68 $ python <path_to_installed_cmdstanpy>/install_cmdstan.py --dir <path_to_build_cmdstan> -v <version_to_use> $ CMDSTAN=<path_to_build_cmdstan>/cmdstan-<version_to_use> STAN_BACKEND=CMDSTANPY pip install prophet==1.0.1
Windows Additional Requirements & Troubleshooting#
If you are using
pip to install EvalML on Windows, it is recommended you first install the following packages using conda:
shapand prediction explanations). Install with
conda install -c conda-forge numba
graphvizif you’re using EvalML’s plotting utilities. Install with
conda install -c conda-forge python-graphviz
The XGBoost library may not be pip-installable in some Windows environments. If you are encountering installation issues, please try installing XGBoost from Github before installing EvalML or install evalml with conda.
Mac Additional Requirements & Troubleshooting#
brew install libomp
graphviz can be installed by running:
brew install graphviz
Installing EvalML on an M1 Mac#
Not all of EvalML’s dependencies support Apple’s new M1 chip. For this reason,
conda installing EvalML will
fail. The core set of EvalML dependencies can be installed in the M1 chip, so we recommend you install EvalML with core
Alternatively, there is experimental support for M1 chips with the Rosetta terminal. After setting up a Rosetta terminal, you should be able to
conda install EvalML.
For Docker fans, an included
Dockerfile.arm can be built and run to provide an environment for testing. Details are included within.