[baymax591]

What does this PR do?

as title

Checklist Before Starting

• Search for similar PRs. Paste at least one query link here: ...
• Format the PR title as [{modules}] {type}: {description} (This will be checked by the CI)
  • {modules} include fsdp, megatron, sglang, vllm, rollout, trainer, ci, training_utils, recipe, hardware, deployment, ray, worker, single_controller, misc, perf, model, algo, env, tool, ckpt, doc, data
  • If this PR involves multiple modules, separate them with , like [megatron, fsdp, doc]
  • {type} is in feat, fix, refactor, chore, test
  • If this PR breaks any API (CLI arguments, config, function signature, etc.), add [BREAKING] to the beginning of the title.
  • Example: [BREAKING][fsdp, megatron] feat: dynamic batching

Test

For changes that can not be tested by CI (e.g., algorithm implementation, new model support), validate by experiment(s) and show results like training curve plots, evaluation results, etc.

API and Usage Example

Demonstrate how the API changes if any, and provide usage example(s) if possible.

# Add code snippet or script demonstrating how to use this

Design & Code Changes

Demonstrate the high-level design if this PR is complex, and list the specific changes.

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 fixes a bug where the TRANSFER_QUEUE_ENABLE setting was not correctly passed to the Ray runtime environment. The change correctly moves this setting into runtime_env.env_vars, which is the correct approach.

My main feedback is regarding significant code duplication between recipe/transfer_queue/main_ppo.py and verl/trainer/main_ppo.py. The run_ppo function is duplicated, and this PR has to apply the same fix in both places. I've left a comment with a suggestion to refactor this to improve maintainability.

[gemini-code-assist]

While this change correctly fixes the bug by setting the TRANSFER_QUEUE_ENABLE environment variable in the Ray runtime environment, it highlights a significant code duplication issue. The run_ppo function in this file is nearly identical to the one in verl/trainer/main_ppo.py, and this PR applies the same fix to both.

This duplication is problematic for maintainability, as any future changes to this logic will need to be applied in both places, increasing the risk of inconsistencies and bugs.

I strongly recommend refactoring to eliminate this duplication. The run_ppo function in this file could be removed. Instead, the main function in this file should import and call run_ppo from verl/trainer/main_ppo.py, passing this recipe's specific TaskRunner as the task_runner_class argument. This would make the code more modular and easier to maintain.