none
What is the difference between Azure Machine Learning Studio and Azure Machine Learning service? What is the intended difference? RRS feed

  • Question

  • What is the difference between Azure Machine Learning Studio and Azure Machine Learning service? What is the intended difference? 

    Despite the ways to train and deploy model(typing code vs. drag and drop), their abilities to handle ML problems seem the same.

    I'm confused to choose one from these two products.

    By the way, does Studio support Deep Learning Development? And in which module?

    Thanks for any reply!

    Thursday, November 1, 2018 2:04 PM

All replies

  • This doc explains the difference between the two offerings. https://docs.microsoft.com/en-us/azure/machine-learning/service/overview-what-is-azure-ml#how-is-azure-machine-learning-service-different-from-studio

    Please let us know if you have any further questions

    Thursday, November 1, 2018 2:52 PM
  • I have the same question, however, the link that you shared, doesn't clarify in depth. 
    Thursday, April 4, 2019 12:05 AM
  • Hey, these are some differences.

    Azure Machine Learning Service(SDK)
    1. Code based Machine Learning experiments for more powerful Customizations using dev tools like Jupyter Notebooks, VisualStudioCode or Azure Notebooks.
    2. Local development possible whereas AzureNotebooks(jupyter on Azure with preinstalled Azure ML sdk) is also available
    3. Python, Azure CLI
    4. Central management of scripts and run history, making it easy to compare model versions(Experiments and Run logs).
    5. SDK supports model management,pipelines, compute management,image creation, deployment etc

    Azure Machine Learning Studio
    1. A Drag-drop style Interactive easy GUI enables machine learning modeling with minimal code.
    2. Online only. No offline development environment
    3. Python,R 
    4. Pick from the available set of algorithms or statistical methods defined as shapes.Though a script shape for R/Python is available.
    5. Use when models are to be created with not much coding involved in less time

    According to me, when a not-so-extensive coder wants to use machine learning, Azure ML Studio will be a good start. But when you want to have more control and wider packages on Python, Azure ML Service is the best place to go.

    This page will help you : https://docs.microsoft.com/en-us/azure/architecture/data-guide/technology-choices/data-science-and-machine-learning

    Thanks


    Sriram N http://srirambiztalks.wordpress.com/


    Thursday, April 4, 2019 3:01 PM