[docs] Example serve llm deployment

[Aydin-ab]

Why are these changes needed?

Adding new ray serve llm deployment examples for both ray docs and Anyscale templates:

• small size llm (single node, 1 gpu) + also show a gated model example with llama 3.1
• medium size llm with tensor parallelism (single node, 2-8 gpus) + also show a gated model example with llama 3.1
• vision llm (sending images, limit # of images per prompt, link to vllm multimodal docs)
• reasoning llm (setup a reasoning parser in LLMConfig, how to parse reasoning vs main content)
• hybrid reasoning llm (setup a reasoning parser in LLMConfig, turn on/off reasoning in a request)

• Also introduce deployment in Anyscale Service, but not too much to not turn off OSS readers

Checks

• I've signed off every commit(by using the -s flag, i.e., git commit -s) in this PR.
• I've run scripts/format.sh to lint the changes in this PR.
• I've included any doc changes needed for https://docs.ray.io/en/master/.
  • I've added any new APIs to the API Reference. For example, if I added a
    method in Tune, I've added it in doc/source/tune/api/ under the
    corresponding .rst file.
• I've made sure the tests are passing. Note that there might be a few flaky tests, see the recent failures at https://flakey-tests.ray.io/
• Testing Strategy
  • Unit tests
  • Release tests
  • This PR is not tested :(

[gemini-code-assist]

Code Review

This pull request adds a comprehensive set of examples for deploying various types of LLMs with Ray Serve, which is a fantastic addition to the documentation. The examples cover different model sizes and functionalities, providing valuable guidance. My review includes several suggestions to fix broken links, correct inconsistencies in model and file names, and resolve some code errors that would prevent the examples from running. Addressing these points will significantly improve the clarity and correctness of the new documentation.

[Aydin-ab]

@khluu
Hi Kevin, FYI, I deleted the old Bazel in the tutorials/ folder as it didn't seem to do anything. It takes precedence over the root BUILD.bazel which might cause confusion for future changes here

[khluu]

stamped.. that BUILD file was modified a long time ago and doesn't seem to be used anywhere in the codebase