Feat: Update RL on Multi-Host TPUs tutorial for clarity and structure

[RexBearIU]

Description

This update reorganizes the multi-host TPU reinforcement learning tutorial for MaxText, Tunix, and vLLM, adding a table of contents and revising the sections for environment setup, checkpoint conversion, and Docker image creation. It separates the steps for stable versus local builds, updates the workload submission commands for GRPO and GSPO, and adds a section for troubleshooting.

Tests

Verified the updated documentation by walking through the entire workflow, including environment setup, Docker image builds, and workload submission. The commands executed successfully as described. Attached are two test logs confirming the results.

• GRPO log

• GSPO log

Checklist

Before submitting this PR, please make sure (put X in square brackets):

• I have performed a self-review of my code. For an optional AI review, add the gemini-review label.
• I have necessary comments in my code, particularly in hard-to-understand areas.
• I have run end-to-end tests tests and provided workload links above if applicable.
• I have made or will make corresponding changes to the doc if needed, including adding new documentation pages to the relevant Table of Contents (toctree directive) as explained in our documentation.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[A9isha]

Approval pending resolution of all comments

Thank you Jacky!