MLA-C01 exam dumps

MLA-C01 practice question 275 of 458

AWS Certified Machine Learning Engineer - Associate. Associate level, Amazon Web Services. Free question with the correct answer and a full explanation.

MLA-C01 Question 275

Single answer

You are tasked with deploying a machine learning model for real-time inference using Amazon SageMaker. The model has specific dependencies that are not included in the pre-built Amazon SageMaker containers. Additionally, you need to optimize the container for faster inference times. How should you choose the appropriate container for this use case?

  1. A

    Use a pre-built SageMaker container and modify its configuration during deployment.

  2. B

    Build a custom container with the required dependencies and performance optimizations.

  3. C

    Use a pre-built SageMaker container and install the required dependencies at runtime.

  4. D

    Build a custom container only if the pre-built containers fail to load your model.

Show answer and explanation

Correct answer: B

Explanation

When deploying a machine learning model with specific dependencies and performance requirements, using a custom container is the best option. A custom container ensures all dependencies are installed and allows for environment optimizations that are crucial for achieving faster inference times. Pre-built SageMaker containers are suitable for general use cases but lack flexibility for advanced scenarios like this one.

  • A. Incorrect.

    This option is incorrect because pre-built SageMaker containers are not designed to be modified extensively during deployment. They are suitable for standard use cases where additional dependencies are not required.

  • B. Correct.

    This is the correct answer because building a custom container allows you to include all necessary dependencies and optimize the environment for your specific model requirements, including performance improvements.

  • C. Incorrect.

    This option is incorrect because installing dependencies at runtime can increase latency and is not an efficient approach for real-time inference scenarios.

  • D. Incorrect.

    This option is incorrect because waiting for the pre-built containers to fail is not a proactive or efficient strategy. Custom containers should be used when you know in advance that pre-built ones do not meet your requirements.

Timed practice exam

Take a MLA-C01 practice test under exam conditions

65 questions in 130 minutes, drawn from this bank, with a score report and a per-question review when you finish.

Start timed exam