A Machine Learning Specialist previously trained a logistic regression model using scikit-learn on a local machine, and the Specialist now wants to deploy it to production for inference only. What…

Question

Options

• ABuild the Docker image with the inference code.
• BSerialize the trained model so the format is compressed for deployment.
• CSerialize the trained model so the format is compressed for deployment.
• DBuild the Docker image with the inference code.

• A
  70% (33)
• B
  4% (2)
• C
  9% (4)
• D
  17% (8)

Explanation

To deploy a locally trained scikit-learn model to Amazon SageMaker for inference, you must bring your own container (BYOC). This requires building a Docker image that includes your inference code (a serving script that loads the model and handles prediction requests). SageMaker expects a specific container interface - the container must respond to /ping and /invocations HTTP endpoints. You also need to serialize the model (e.g., via joblib or pickle), upload the artifact to S3, and register a SageMaker Model pointing to both the Docker image and the S3 artifact. The question's options A and D are identical, as are B and C - a formatting artifact. The key step distinguishing deployment readiness is building the Docker image with inference code, because SageMaker cannot natively run a raw scikit-learn model object without a properly configured serving container.

No community discussion yet for this question.