ML.NET 1.4 Preview Released

ML.NET 1.4 Preview is now available with new features like Database-Loader and Image classification with deep neural networks retraining.

Recently, Microsoft announced ML.NET 1.4 Preview and new updates to Model Builder and CLI. The 1.4 Preview is a backward-compatible release with no breaking changes. It includes a bunch of bug fixes and some exciting new features.
 
Well, ML.NET is an open-source and cross-platform machine learning framework from Microsoft. It includes Model Builder, which is a simple UI tool, and CLI. ML.NET tools make it very easy to build custom ML models using Automated Machine Learning (AutoML).
 
ML.NET 1.4 Preview brings a new feature Database-Loader(Preview) which introduces a native database loader that enables training directly against relational databases. You can use any RDBMS like SQL Server, Azure SQL Database, Oracle, SQLite, PostgreSQL, etc.
 
database_loader 
Source: Microsft 
 
Interestingly, this new Database Loader provides a way easier code implementation. Because the way it reads from the database and makes data available through the IDataView is provided out-of-the-box by the ML.NET framework. You just have to specify your database connection string, the SQL statement for the dataset columns and the data-class to use.
 
The new release also features 'Image classification with deep neural networks retraining (Preview)', which is new feature that enables native DNN transfer learning with ML.NET, targeting image classification as the first high-level scenario. For example, you can now create your own custom image classifier model by natively training a TensorFlow model from ML.NET API with your own images.
 
 ML.NET IMAGE CLASSIFIER
Microsoft said that ML.NET is now building for .NET Core 3.0. You can take advantage of the new features when running in a .NET Core 3.0 application. The new hardware intrinsics feature allows .NET code to accelerate math operations by using processor-specific instructions.
 
So, when training on a modern processor, ML.NET will now train much faster. Because it can do more concurrent floating-point operations than it could with the existing C++ code that only supported SSE instructions
 
Microsoft also announced that the Model Builder tool in Visual Studio and the ML.NET CLI (both in preview) have been updated to use the latest ML.NET GA version (1.3). This will allow developers who want to use their own local culture OS settings to train a model in Model Builder.
 
For those who want to learn ML.NET, the team has also created a ML.NET Youtube playlist with each video focusing on a single and particular ML.NET feature.