70-774 · Question #59
You are building an Azure Machine Learning workflow by using Azure Machine Learning Studio. You create an Azure notebook that supports the Microsoft Cognitive Toolkit. You need to ensure that the…
The correct answer is A. DataParallelASGD. Data-Parallel Training with Parameter Server Parameter server is a widely used framework in distributed machine learning. The most important benefit it brings is the asynchronous parallel training with many workers. You can use Data-Parallel ASGD. Incorrect ANswers: D…
Question
You are building an Azure Machine Learning workflow by using Azure Machine Learning Studio. You create an Azure notebook that supports the Microsoft Cognitive Toolkit. You need to ensure that the stochastic gradient descent (SGD) configuration maximizes the samples per second and supports parallel modeling that is managed by a parameter server. Which SGD algorithm should you use?
Options
- ADataParallelASGD
- BDataParallelSGD
- CBlockMomentumSGD
- DModelAveragingSGD
How the community answered
(25 responses)- A80% (20)
- B12% (3)
- C4% (1)
- D4% (1)
Explanation
Data-Parallel Training with Parameter Server Parameter server is a widely used framework in distributed machine learning. The most important benefit it brings is the asynchronous parallel training with many workers. You can use Data-Parallel ASGD. Incorrect ANswers: D: Model-Averaging SGD generally converges more slowly and to a worse optimum, compared to 1-bit SGD and Block-Momentum SGD, so it is no longer recommended. https://docs.microsoft.com/en-us/cognitive-toolkit/multiple-gpus-and-machines
Topics
Community Discussion
No community discussion yet for this question.