[Docs] Add TensorRT Model Connect to the TensorRT documentation

Summary

Add TensorRT Model Connect to the main TensorRT documentation so users can discover the simplest path from a supported PyTorch model to end-to-end TensorRT inference without already knowing the project name.

The documentation should explain when Model Connect is the appropriate entry point and lead users directly to a working quickstart.

Content

• Explain what TensorRT Model Connect provides and who should use it.
• Include the two-command build-and-run workflow:

trtmc build Qwen/Qwen3-0.6B --max-cache-length 16384 --output qwen3-0.6b.bundle
trtmc run ./qwen3-0.6b.bundle --prompt "What is the capital of France? Answer in one word." --chat-template --no-thinking
# Generated text: Paris

• Explain the relationship between the three layers:
  • TensorRT provides the underlying inference optimizer and runtime.
  • TensorRT Model Connect provides model-level reference implementations, bundle generation, and task-oriented native runtime workflows.
  • TensorRT Edge-LLM is the production-focused path for qualified LLM and VLM deployment on supported NVIDIA edge platforms.
• Link to:
  • Model Connect quickstart
  • Installation and environment setup
  • Supported models and qualification profiles
  • Tutorials
  • GitHub repository
• Clearly distinguish qualified support for an exact checkpoint and configuration from best-effort compatibility with related checkpoints or configurations.
• Document the relevant checkpoint revision, precision, platform, and runtime-version boundaries where support claims are made.
• Add Model Connect to the relevant navigation, ecosystem, and getting-started surfaces in the main TensorRT documentation.
• Ensure the TensorRT documentation directs eligible PyTorch-model users to the Model Connect quickstart.

Acceptance criteria

A user entering through the TensorRT documentation can:

• Understand what TensorRT Model Connect provides.
• Decide when to use Model Connect, TensorRT Edge-LLM, or lower-level TensorRT APIs.
• Reach and complete a working Model Connect quickstart without already knowing the project name.
• Distinguish verified configurations from best-effort compatibility.

Start with the main TensorRT documentation, its navigation, ecosystem, and getting-started surfaces, then review the linked Model Connect quickstart, setup overview, supported-models page, tutorials, and repository. Add the requested Model Connect guidance, commands, links, support boundaries, and routing for eligible PyTorch users; done means a reader can choose the appropriate TensorRT path and complete the quickstart.