A harbor-radar team needs a custom attention loop that no SageMaker built-in ships. They already write PyTorch and do not need a new AWS service. Which training path fits?
Select an answer to reveal the explanation.
Short Explanation
Think of a custom attention loop that no built-in ships, and a team that already writes PyTorch. SageMaker script mode with the PyTorch estimator. Do not force a built-in that cannot run that loop.
Full Explanation
SageMaker script mode with the PyTorch estimator and container is the path for custom training code in a supported framework. A built-in cannot run an attention loop it does not ship. Transcribe and Fraud Detector are not script-mode trainers.