Pass SIGTERM to training script to stop training

**Describe the bug**
SIGTERM from StopTrainingJob doesn't appear to be passed to the training subprocess.

**To reproduce**
Add a SIGTERM handler to a training script, start a training job, then click "Stop". The signal handler will not fire.

**Expected behavior**
Signal handler should fire when "StopTrainingJob" happens

**Screenshots or logs**
If applicable, add screenshots or logs to help explain your problem.

**System information**
A description of your system.
- Include the version of SageMaker Training Toolkit you are using.
- If you are using a [prebuilt Amazon SageMaker Docker image](https://docs.aws.amazon.com/sagemaker/latest/dg/pre-built-containers-frameworks-deep-learning.html), provide the URL.
- If you are using a custom Docker image, provide:
- framework name (eg. PyTorch)
- framework version
- Python version
- processing unit type (ie. CPU or GPU)

**Additional context**
Add any other context about the problem here.

Reproduce the issue by running a training job with a SIGTERM handler and stopping it through StopTrainingJob. Locate the training subprocess entry point and trace how the stop signal reaches it; the work is done when the handler fires after clicking Stop and the behavior is covered by an appropriate regression test.

Written by the indexing model from the issue text.