Revise introduction and usage details for DLCs

[jeffboudier]

Updated introduction to clarify the role of Hugging Face and AWS in maintaining Deep Learning Containers. Enhanced descriptions of usage and features across various AWS services. Removed references to TGI in favor of vLLM, SGLang, llama.cpp

---

Note

Low Risk
Documentation-only edits to a single markdown file with no runtime, API, or security impact.

Overview
Rewrites docs/sagemaker/source/dlcs/introduction.md to match current DLC positioning and the inference lineup documented on the Available DLCs page.

The intro now states that Hugging Face and AWS jointly maintain DLCs, points readers to ECR and the Available DLCs doc, and narrows the AWS services list to SageMaker AI instead of Bedrock, EKS, ECS, and EC2.

High-performance text generation no longer centers TGI; it documents dedicated DLCs for vLLM, SGLang, and llama.cpp, plus OpenAI-compatible APIs and S3 model loading. TEI gets its own High-performance embeddings section. Section titles and copy are tightened throughout (training TRL CLI, inference toolkit, built-in performance).

^{Reviewed by Cursor Bugbot for commit 6d4edfb. 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.