MLflow (machine learning of a full stroke)

2020: Zapuks project on implementation of "turnkey" machine learning

At the end of June, 2020 announced Linux Foundation start under the wing of the MLflow project within which the platform for implementation of a full stroke of machine learning develops of the same name.

The MLflow project which was originally developed by Databricks company simplifies creation and deployment of models of machine learning and also exchange of them. The tool assumes the API set which work with any library, including with PyTorch, TensorFlow, and XGBoost and also in any environment, including cloud services.

MLflow is a project for those developers who operate with a huge number of different models, with different bases of the code, diagrams, metrics and need the tool which allows to trace all indicators at code optimization and setup of model for further development. At MLflow is built in integration with Docker, TensorFlow, PyTorch, Kubernetes, Java, Spark and other open projects. This tool is already used by such companies as Microsoft, Accenture, SK Telecom and also the University of Washington.

Linux Foundation started the project on implementation of "turnkey" machine learning

The founder of MLflow Matey Zachariah noted that this tool became the standard for platforms of machine learning thanks to community of participants which consists of hundreds of engineers from more than hundred companies. Machine learning continues to transform processes in all industries, allowing to carry out billions of tasks automatically.

It represents the special importance for health care, finance and retail. Partnership of MLflow and Linux Foundation will allow to involve in use of a versatile tool for creation and testing of models of machine learning even more developers. The developers interested in the tool can find more information on how to set and use MLflow in the blog of the project on GitHub.^[1][2]

Notes