Add example: deploy Faster-Whisper as a custom model on a SageMaker Triton MME

[eldadcohen1]

Adds an example showing how to deploy a pre-trained Faster-Whisper-Large-v3 model from Hugging Face as a custom model on an Amazon SageMaker multi-model endpoint (MME) backed by Triton, including building a custom Triton container and running audio-transcription inference.

Description of changes:

• New example under deploy_and_monitor/sm-triton_mme_faster_whisper/ (notebook, README.md, and a sample audio file), following the current repository structure and sm-triton_* naming convention.
• Registered the notebook in the Deploy and Monitor toctree in index.rst.
• CI badge URLs and kernelspec (conda_python3) match the sibling Triton examples.

This supersedes #4608 / #4887, which targeted the legacy main branch and the old sagemaker-triton/ layout. This PR targets the current default branch and the new folder structure.

Testing is done:
Tested on AWS SageMaker.

Merge Checklist

• I have read the CONTRIBUTING doc and adhered to the example notebook best practices
• I have updated any necessary documentation, including READMEs
• I have tested my notebook(s) and ensured it runs end-to-end
• I have linted my notebook(s) and code using black-nb -l 100 {path}/{notebook-name}.ipynb

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[eldadcohen1]

@mollyheamazon can you please review and approve

[mollyheamazon]

The notebook uses backend: "python" in config.pbtxt, but there's no cell that writes the required model.py into the model directory. Could you add the %%writefile cell for it? Without it a reader can't reproduce end-to-end.

Another note: This notebook uses raw boto3 calls rather than the SageMaker Python SDK. It might be a better fit for awsdocs/aws-doc-sdk-examples (which hosts boto3 examples)

[eldadcohen1]

Thanks @mollyheamazon!

1. Missing model.py - good catch, fixed in 6d68871. Added a %%writefile faster-whisper-large-v3/1/model.py cell that writes the Python-backend model into the version directory next to the weights. It loads WhisperModel in initialize, and in execute base64-decodes audio_array, transcribes with the given audio_lang, and returns transcript. faster-whisper is already pip-installed in the Dockerfile, so the example now reproduces end-to-end.

2. boto3 vs SageMaker SDK - the raw boto3 calls are intentional here: the example's focus is the low-level MME + custom-Triton-container flow (create_model / create_endpoint_config / create_endpoint with Mode: MultiModel), which the SDK abstracts away. Happy to defer to you, though - if you'd rather this live in aws-doc-sdk-examples, let me know and I'll move it.

[eldadcohen1]

@jam-jee , @mollyheamazon fixed you comments, can you please take a look.

[eldadcohen1]

@mollyheamazon any thoughts?