add: tutorial about multilora using inference components

[dwarez]

Note

Low Risk
Documentation-only addition (a Jupyter notebook); no application or library code changes.

Overview
Adds a new SageMaker SDK notebook at docs/sagemaker/notebooks/sagemaker-sdk/multi-lora-adapters-vllm/sagemaker-notebook.ipynb that walks through serving multiple LoRA adapters from one GPU endpoint using the Hugging Face vLLM DLC and SageMaker inference components.

The tutorial covers packaging Hub adapters into S3 tar.gz archives, enabling LoRA and runtime adapter updates via vLLM env vars (including VLLM_ALLOW_RUNTIME_LORA_UPDATING), deploying Mistral 7B as a base inference component with SageMaker Python SDK v3 (Model, Endpoint, InferenceComponent), attaching adapter components, routing with InferenceComponentName and route=/v1/completions, and adding/removing adapters without redeploying. It also includes production notes and a cleanup section for billable resources.

^{Reviewed by Cursor Bugbot for commit 9c392c8. Bugbot is set up for automated code reviews on this repo. Configure here.}

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 819b995. Configure here.}