MLS-C01 · Question #34
A Machine Learning team has several large CSV datasets in Amazon S3. Historically, models built with the Amazon SageMaker Linear Learner algorithm have taken hours to train on similar- sized…
The correct answer is A. Use Amazon SageMaker Pipe mode. Amazon SageMaker Pipe mode streams training data directly from S3 to the training algorithm as it is needed, rather than first copying the entire dataset to the training instance's local EBS volume (which is what the default File mode does). For large datasets, this eliminates…
Question
A Machine Learning team has several large CSV datasets in Amazon S3. Historically, models built with the Amazon SageMaker Linear Learner algorithm have taken hours to train on similar- sized datasets. The team's leaders need to accelerate the training process. What can a Machine Learning Specialist do to address this concern?
Options
- AUse Amazon SageMaker Pipe mode.
- BUse Amazon Machine Learning to train the models.
- CUse Amazon Kinesis to stream the data to Amazon SageMaker.
- DUse AWS Glue to transform the CSV dataset to the JSON format.
How the community answered
(27 responses)- A81% (22)
- B4% (1)
- C4% (1)
- D11% (3)
Explanation
Amazon SageMaker Pipe mode streams training data directly from S3 to the training algorithm as it is needed, rather than first copying the entire dataset to the training instance's local EBS volume (which is what the default File mode does). For large datasets, this eliminates the lengthy data-download phase before training even starts, dramatically reducing total training time. Pipe mode also reduces storage requirements on the instance. Option B (Amazon Machine Learning) is a legacy service with limited algorithm support and is not a drop-in replacement. Option C (Kinesis streaming) adds unnecessary infrastructure complexity. Option D (converting CSV to JSON with Glue) does not accelerate training and JSON is not a more efficient format for SageMaker's Linear Learner; RecordIO-protobuf would be better.
Topics
Community Discussion
No community discussion yet for this question.