MLA-C01 Question 435
Select 2You are developing a machine learning model to classify images of vehicles into categories such as 'car', 'truck', and 'motorcycle'. The dataset is stored in an Amazon S3 bucket. You want to preprocess the data using AWS Glue and train the model using Amazon SageMaker. After training, you need to ensure the model achieves high accuracy by optimizing its hyperparameters. Which combination of steps should you take to achieve this goal?
- A
Use AWS Glue to clean and transform the dataset, and store the processed data back into the S3 bucket.
- B
Use Amazon SageMaker's built-in algorithms to train the model and manually select hyperparameters.
- C
Enable Amazon SageMaker Automatic Model Tuning to optimize hyperparameters.
- D
Store the training and validation datasets in Amazon RDS for increased performance during training.
- E
Leverage Amazon SageMaker Data Wrangler to preprocess the data instead of AWS Glue.
Show answer and explanation
Correct answers: A, C
Explanation
To preprocess the dataset, AWS Glue is a scalable and efficient choice for cleaning and transforming large datasets stored in Amazon S3. After preprocessing, Amazon SageMaker Automatic Model Tuning should be used to optimize hyperparameters systematically, ensuring the model achieves high accuracy. Manual hyperparameter tuning is less efficient, and storing datasets in RDS is unnecessary for this workflow. While Data Wrangler is an alternative preprocessing tool, AWS Glue is better suited for large-scale data transformation in this case.
- A. Correct.
Correct: AWS Glue is well-suited for cleaning and transforming large datasets stored in Amazon S3. It integrates well with SageMaker and is often used in ML workflows.
- B. Incorrect.
Incorrect: While SageMaker's built-in algorithms are useful, manually selecting hyperparameters is less efficient and prone to error compared to using automated tuning tools like Amazon SageMaker Automatic Model Tuning.
- C. Correct.
Correct: Amazon SageMaker Automatic Model Tuning is the preferred method to systematically optimize hyperparameters and achieve better performance for the ML model.
- D. Incorrect.
Incorrect: Storing datasets in Amazon RDS is not ideal for ML workflows. S3 is the recommended choice as it is cost-effective, scalable, and directly integrates with SageMaker.
- E. Incorrect.
Incorrect: While Amazon SageMaker Data Wrangler is powerful for preprocessing, in this scenario, AWS Glue is already chosen and is more suitable for large-scale transformation and cleaning tasks.