News

Microsoft Updates ML.NET, Offers Personal Help for Production Use

• By David Ramel
• 03/14/2019

Microsoft updated its open source machine learning framework, ML.NET, to version 0.11 and offered up an engineer to help developers make it work in a production environment.

ML.NET was announced at the company's Build developer conference last May. Microsoft said it helps developers apply their .NET and C# or F# skills to integrate custom machine learning into an application even if such developers don't have expertise in developing or tuning machine learning models.

Such models can be used for projects needing functionality such as sentiment analysis, recommendation, image classification and so on.

While ML.NET 0.11 was described as a stability-focused release, the update comes with a twist: Microsoft is offering to provide one-on-one help from an engineer to help developers use it in production.

The company said in a post last week: "If you are using ML.NET in your app and looking to go into production, you can talk to an engineer on the ML.NET team to:

• Get help implementing ML.NET successfully in your application.
• Provide feedback about ML.NET.
• Demo your app and potentially have it featured on the ML.NET homepage, .NET Blog, or other Microsoft channel.

Microsoft provided a form to fill out for that feedback mechanism, which asks a series of questions and lets developers leave their contact information if they want someone from the ML.NET team to contact them.

In the meantime, the team said all future releases leading up to version 1.0 will focus on stability, with API refinements, bug fixes, reduction of the public API surface, improved documentation and samples.

Two notable highlights of the 0.11 update were listed as: new capabilities to support text input when working with the popular TensorFlow, which facilitates working with text analysis scenarios such as sentiment analysis; some renaming of ONNX-related terms "to make the distinction between ONNX conversion and transformation clearer." ONNX is an open and interoperable model format that lets developers take models trained in one ML framework and use them in another framework.