ML.NET is a cross-platform and open-source machine learning framework that consists of CLI (Command-Line Interface) and Model Builder (which is a simple UI tool ) to make it easy to build custom ML models using Automated Machine Learning (AutoML).
Using ML.NET, developers can integrate ML.NET with their existing tools to develop and infuse their applications with custom ML by creating custom ML models for some common scenarios like Image Classification, Sentiment Analysis, Sales Forecast prediction, Price Prediction, Customer segmentation, and more!
What’s new in ML.NET 1.4 –
- ML.NET 1.4 enables image classification based on Deep Neural Network retraining with GPU support (GA release).
Here are models with GPU support:
ML.NET 1.4 brings the feature of Database-Loader(Preview) which introduces a native database loader that enables training directly against relational databases. You can use any RDBMS like Azure SQL Database, SQL Server, Oracle, SQLite, PostgreSQL, etc.
In the new Database Loader code implementation is way too easy. Because the way it reads from the database and makes the data available through the IDataView is provided exceptionally by the ML.NET framework. Just the database connection string has to be specified, the SQL statement for the dataset columns, and the data-class to use.
The new release also features ‘Image classification with DNN retraining (Preview)’, which is a new feature that enables native DNN transfer learning with ML.NET, targeting image classification as the first high-level scenario.
For example, it is easy for us to create our own custom image classifier model by natively training a TensorFlow model from ML.NET API with our own images.
ML.NET is now building for .NET Core 3.0.
So, while training on a modern processor, ML.NET is now trained much faster because of the ability to do more concurrent floating-point operations than it could with the existing C++ code that only supported SSE instructions
Microsoft also announced about 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). It will help the developers who want to use their own local culture OS settings to train a model in the Model Builder.
The ML.NET team has also created an ML.NET Youtube playlist with each video focusing on a single and particular ML.NET feature.