Exam DP-100 topic 2 question 4 discussion

It seems like the answer explanation is mixing Azure Machine Learning Studio with Azure Machine Learning Designer since the description is for designer not studio. Studio includes designer, notebook and autoML. I think the correct answer is actually B since Azure Machine Learning Service can only be used from within Azure Machine Learning Studio.

I just did some digging and here is what I find: https://docs.microsoft.com/en-us/azure/machine-learning/overview-what-is-machine-learning-studio. The "Machine Learning Studio" in this question is actually referring to "Machine Learning Studio Classic", which is an outdated platform that only offering web service (no offline), and the "Machine Learning Service" is actually the "Azure Machine Learning" you can find in Azure Portal. What's the more the new Studio is now part of the "Azure Machine Learning" as we speak. So for this question, A is the correct answer at the time when it was firstly created, but it no long applies right now. Now, A and B are pretty much the same thing.

A. Azure Machine Learning Service Supports large datasets and custom frameworks (like Caffe2, Chainer via Docker). Allows local development and offline work using Azure ML SDK and local compute. Pipelines can be updated when reconnected. ✅ Best fit for all requirements.

Azure Machine Learning Service is the most appropriate choice based on the requirements: Supports Caffe2 and Chainer frameworks: Azure Machine Learning Service allows flexibility in choosing machine learning frameworks, including Caffe2 and Chainer, through custom environments. Supports both connected and disconnected environments: With Azure Machine Learning Service, data scientists can work locally on their personal devices using the Azure ML SDK or CLI, even offline. Pipelines and models can be updated and synced when reconnected to the network. Large dataset support: The service can handle datasets larger than 20 GB by enabling integration with cloud storage (e.g., Azure Blob Storage) and local caching mechanisms for disconnected environments. Personal device flexibility: Azure Machine Learning Service supports local development and allows deploying models on personal devices with proper synchronization capabilities. Chat GPT

ChatGPT says D.

A is correct! Azure Machine Learning Service supports local machine compute!

https://www.codit.eu/blog/azure-machine-learning-studio-vs-services/

A. Azure Machine Learning Service would be the best option for building a team data science environment with the given requirements. It allows building machine learning pipelines using Caffe2 or Chainer frameworks, supports training models on personal devices in both connected and disconnected network environments, and provides a mechanism for updating machine learning pipelines when connected to a network.

on this link, I found the difference between azure ML service and studio and the conclusion is azure Ml service is hybrid(on cloud or premise) https://www.codit.eu/blog/azure-machine-learning-studio-vs-services/?country_sel=uk

C is correct

Option A

option A

Answer is A. Because Azure Machine Learning Studio need network! Azure Machine Learning Service can support local compute. https://www.codit.eu/blog/azure-machine-learning-studio-vs-services/?country_sel=be

Why is B incorrect??

answer should be A, but already out-dated ? https://www.codit.eu/blog/azure-machine-learning-studio-vs-services/?country_sel=be

What's the consensus on the answer? Is it A or B? I cannot find any learning document that provides clarity on compatibility with caffe2 or chainer, and any information on off network local machine usage.

I think the answer is A, Azure ML service is hybrid (can be run both on cloud and on premise) https://www.codit.eu/blog/azure-machine-learning-studio-vs-services/