Reorganization of ToC

[mtsokol]

Hi @RobMulla,

Here's a new structure of the docs for vLLM-TPU. It's divided into three main sections:

• Getting Started - a minimal set of actions to run an inference tasks with vLLM on a TPU.
• Basic Usage - for users with supported HF models, deploying it on a single TPU manually, or GCE cluster.
• Advanced Usage - for power users / developers wanting to run inference on a custom JAX/Torch models on a single TPU or GCE.

It's proposed based on this tree:

[RobMulla]

Thanks @mtsokol for your work on this!

@bvrockwell and @ica-chao - these are some of the docs re-organization we are working on to streaming UX. Separate than our support matrix work. When you have a chance could you review?

Rendered docs can be found here (or the link in the CI/CD of this PR) - https://vllm--2064.org.readthedocs.build/projects/tpu/en/2064/ and you can toggle the diff to see whats changed.