[recipe, hardware] feat: Add GRPO with full weight updates for 1.5B models on a single GPU

[KabakaWilliam]

What does this PR do?

Adds a new recipe to perform GRPO with full weight-updates on 1.5B models.

Checklist Before Starting

• Search for similar PRs. Paste at least one query link here: [recipe] fix: make LangGraph agent example runnable out-of-the-box #3029
• Format the PR title as [recipe, hardware] feat: Add GRPO with full weight updates for 1.5B models on a single GPU (This will be checked by the CI)
  • {modules}: recipe, hardware
  • {type}: feat
  • No breaking API changes

Test

> Pass@1 for MATH500: 77.8

Checklist Before Submitting

Important

Please check all the following items before requesting a review, otherwise the reviewer might deprioritize this PR for review.

• Read the Contribute Guide.
• Apply pre-commit checks: pre-commit install && pre-commit run --all-files --show-diff-on-failure --color=always
• Add / Update the documentation.
• Add unit or end-to-end test(s) to the CI workflow to cover all the code. If not feasible, explain why: ...
• [] Once your PR is ready for CI, send a message in the ci-request channel in the verl Slack workspace. (If not accessible, please try the Feishu group (飞书群).)

[gemini-code-assist]

Code Review

This pull request introduces a new recipe for GRPO with full weight updates on 1.5B models. My review identified two critical issues in the new shell script qwen2.5math-1.5b_grpo_1_h100_fsdp_vllm.sh where missing spaces before line continuation characters would cause the script to fail due to malformed arguments. Additionally, I've noted that the pull request template file has been incorrectly modified with PR-specific details and should be reverted.

[KabakaWilliam]

removed continuation character from end of script and ensured proper spacing, reverted the Pull request templat eback to it's normal state