Attention! This repository is deprecated, unmaintained, and will get no further updates!
This repository hosts Wave ML (h2o-wave-ml
), a companion package for H2O Wave that makes it quick and easy to integrate AI/ML models into your applications.
Wave ML provides a simple, high-level API for training, deploying, scoring and explaining machine learning models, letting you build predictive and decision-support applications entirely in Python.
Wave ML runs on Linux, OSX, and Windows, and utilizes H2O.ai's open-source H2O and AutoML under the hood.
When Wave apps are run in H2O AI Hybrid Cloud with GPU support, Wave ML optionally switches over to Driverless AI for automatic feature engineering, machine learning, model deployment, and monitoring.
Both Wave and Wave ML can be installed in tandem using pip:
> pip install h2o-wave[ml]
You can install Wave ML separately as well:
> pip install h2o-wave-ml
Note that the package pulled from PyPI doesn't have all the dependencies needed for Cloud development. Steam and MLOps packages are missing and have to be installed separately. However, they are present in the package from this repo in release section. Look into setup.py to see the packages and include them in requirements file if needed.
See a quickstart guide here.
The API can be found on official Wave page here.
Check a Wiki for a guide.
H2O Wave ML is licensed under the Apache License 2.0. See LICENSE for more information.