Exam AWS Certified Machine Learning - Specialty topic 1 question 188 discussion

https://docs.aws.amazon.com/sagemaker/latest/dg/your-algorithms-training-algo-dockerfile.html

The /opt/ml directory is the default directory where SageMaker expects the training script and other related files to be located. The script at location above is triggered by setting environment variable SAGEMAKER_PROGRAM and *not* through an ENTRYPOINT in docker file

A. NO - There is no server here, we do training not inference B. YES C. NO - path to training data is externally provided, not hardcoded in the image D. NO - /opt/ml/train is the working directory of the ENTRYPOINT

Amazon SageMaker supports bringing custom training algorithms by using Docker containers, which are software packages that can contain all the dependencies and configurations needed to run an application. Dockerfile is a text file that contains the instructions for building a Docker image, which is a snapshot of a Docker container. ENTRYPOINT is an instruction in the Dockerfile that specifies the default executable or command that will run when the container is started. By specifying the training program in the ENTRYPOINT instruction, the ML specialist can ensure that Amazon SageMaker can run the training program automatically when it creates and runs a Docker container for the training job.

https://docs.aws.amazon.com/sagemaker/latest/dg/adapt-training-container.html In Step 2, it is mentioned to use this instruction on dockerfile: # Defines train.py as script entrypoint ENV SAGEMAKER_PROGRAM train.py