Exam AWS Certified Machine Learning - Specialty topic 1 question 225 discussion

It has to be C.

Option C uses the SageMaker Debugger vanishing_gradient and LowGPUUtilization built-in rules to detect issues and to launch the StopTrainingJob action if issues are detected. This option is the most efficient because it leverages the existing features of SageMaker Debugger to monitor and troubleshoot your training job without requiring any additional development effort. You can use the following steps to implement this option.

Answer is C The best option for the data scientist to identify and address training issues with the least development effort is option C: Use the SageMaker Debugger vanishing_gradient and LowGPUUtilization built-in rules to detect issues and to launch the StopTrainingJob action if issues are detected. SageMaker Debugger is a tool that helps to debug machine learning training processes. It provides several built-in rules that can detect and diagnose common issues that can occur during training. In this case, the data scientist suspects that the training is not converging and that resource utilization is not optimal. The vanishing_gradient and LowGPUUtilization rules can help to identify these issues.

C. Use the SageMaker Debugger vanishing_gradient and LowGPUUtilization built-in rules to detect issues and to launch the StopTrainingJob action if issues are detected. The SageMaker Debugger is a built-in tool that helps with debugging and profiling machine learning models trained in SageMaker. In this scenario, the data scientist suspects that there are issues with the training process, so using the SageMaker Debugger is the most appropriate solution. The vanishing_gradient and LowGPUUtilization built-in rules can detect common training issues such as a vanishing gradient problem or low GPU utilization, which could affect the training convergence and resource utilization. By launching the StopTrainingJob action if issues are detected, the training job can be stopped early, which can help to save resources and time. This approach requires the least development effort, as it is built-in to SageMaker and does not require the data scientist to create custom metrics or configure CloudWatch alarms.

https://docs.aws.amazon.com/sagemaker/latest/dg/debugger-built-in-rules.html