Answer-first summary for fast verification
Answer: No
The correct answer is 'No' (B) because Bayesian sampling in Azure Machine Learning does not support early termination policies. According to the official Microsoft documentation referenced in the community discussion, Bayesian sampling requires using NoTerminationPolicy, setting early termination policy to None, or omitting the early_termination_policy parameter entirely. While Bayesian sampling does support continuous hyperparameters (as mentioned in one comment), it fails to meet the requirement for early termination of low-performance jobs. The community consensus strongly supports answer B (67% vs 33% for A), with the most upvoted comment providing authoritative documentation evidence. For scenarios requiring both early termination and continuous hyperparameters, other sampling methods like Random or Grid sampling with appropriate termination policies would be more suitable.
Author: LeetQuiz Editorial Team
Ultimate access to all questions.
You plan to tune model hyperparameters using a sweep job in your Azure Machine Learning workspace.
You need to select a sampling method that supports both early termination of low-performance runs and the use of continuous hyperparameters.
Proposed solution: Use the Bayesian sampling method over the hyperparameter space.
Does this solution meet the goal?
A
Yes
B
No
No comments yet.