[ETOgaosion]

Reverts #2739

For #2794 to solve all CI faults.

[gemini-code-assist]

Code Review

This pull request reverts a previous feature related to vLLM service sleep/wakeup modes, presumably to fix CI issues. The changes correctly remove the feature's implementation and documentation. My review focuses on the resulting code within the diff. I've identified an opportunity to improve the conciseness and maintainability of the task submission logic in reward_function.py by using a more idiomatic Python construct. The provided suggestion will make the code cleaner without altering its behavior.

[gemini-code-assist]

The for loop used to submit tasks to the ThreadPoolExecutor is unnecessarily verbose. This can be refactored into a more concise and idiomatic list comprehension.

This change improves readability and maintainability by replacing a multi-line loop with a single, expressive statement, which is a common pattern in modern Python for this type of task.

Suggested change

	for data_source, solution_str, ground_truth, extra_info in zip(
	data_sources, solution_strs, ground_truths, extra_infos, strict=True
	):
	future = executor.submit(compute_score, data_source, solution_str, ground_truth, extra_info, index)
	time.sleep(0.001 * random.random())
	future = executor.submit(compute_score, data_source, solution_str, ground_truth, extra_info)
	futures.append(future)
	futures = [
	executor.submit(compute_score, data_source, solution_str, ground_truth, extra_info)
	for data_source, solution_str, ground_truth, extra_info in zip(
	data_sources, solution_strs, ground_truths, extra_infos, strict=True
	)
	]