Programming book reviews, programming tutorials,programming news, C#, Ruby, Python,C, C++, PHP, Visual Basic, Computer book reviews, computer history, programming history, joomla, theory, spreadsheets and more.
Microsoft has released an open-source library for creating massively scalable machine learning (ML) pipelines. SynapseML was until now known as MMLSpark, and it unifies several existing ML frameworks and new Microsoft algorithms in a single, scalable API that’s usable across Python, R, Scala, and Java. SynapseML builds on Apache Spark and SparkML, adding deep learning and data science tools to the Spark ecosystem. It integrates Spark Machine Learning pipelines with the Open Neural Network Exchange (ONNX), LightGBM, The Cognitive Services, Vowpal Wabbit, and OpenCV, so through these tools provides highly-scalable predictive and analytical models for a variety of data sources. SynapseML also includes the HTTP on Spark project, meaning users can embed web services into their SparkML models.
