Feat/ddp mlflow

[KeitaW]

Issue #, if available:

Description of changes:

This PR makes ddp test case compatible with both CPU and GPU + add MLFlow support (local tracking by default with optional remote tracking server option).

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

[KeitaW]

The test case is already compatible with local/remote tracking server.

[allela-roy]

@KeitaW , to showcase MLflow, shouldn't we use the SageMaker Managed MLflow as described here https://docs.aws.amazon.com/sagemaker/latest/dg/mlflow-track-experiments.html

[allela-roy]

Working on the compatibility with Managed MLflow

[KeitaW]

@KeitaW , to showcase MLflow, shouldn't we use the SageMaker Managed MLflow as described here https://docs.aws.amazon.com/sagemaker/latest/dg/mlflow-track-experiments.html

Thanks for testing.
It should work with ManagedMLFlow too. How to set up etc. should be addressed in the architecture sub directory not in the test case (we may add link to the guidance).

[allela-roy]

Ok, then let's add this to pip install mlflow==2.13.2 sagemaker-mlflow==0.1.0 to the 0.create-venv.sh as well.