[sglang] Fixing issues in examples/sglang_multiturn/search_r1_like/run_qwen2.5-3b_instruct_search_multiturn.sh

[NLPJCL]

What does this PR do?

When training search-r1 using Verl, I found that the script wouldn't run. This was probably because subsequent updates to the Verl agent loop didn't have compatible parameters. After fixing it, it started running.

1. To adapt to the tool call format of Verl's agentlloop, the prompt of search-r1 was modified.
2. actor_rollout_ref.rollout.multi_turn.format: Modify the template from non-existent qwen to hermes.
3. actor_rollout_ref.rollout.multi_turn.max_assistant_turns=4; if actor_rollout_ref.rollout.multi_turn.max_assistant_turns set=2,It only performs one search, which is inconsistent with the multi-round search in search-r1.
4. actor_rollout_ref.rollout.multi_turn.max_tool_response_length=1024. Since local search wikis are composed of English words, the length of the output mode is increased to avoid truncation by the default configuration.
   

   verl/verl/experimental/agent_loop/tool_agent_loop.py

   Line 445 in b178a3c

   if tool_response_text and len(tool_response_text) > self.max_tool_response_length:

   
   

   verl/verl/workers/config/rollout.py

   Line 55 in b178a3c

   max_tool_response_length: int = 256

Add concise overview of what this PR aims to achieve or accomplish. Reference related GitHub issues and PRs that help with the review.

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, veomni, sglang, vllm, rollout, trainer, ci, training_utils, recipe, hardware, deployment, ray, worker, single_controller, misc, perf, model, algo, env, tool, ckpt, doc, data, cfg, reward
  • 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).)
• If your PR is related to the recipe submodule, please also update the reference to the submodule commit via git submodule update --remote or cd recipe && git pull origin main.

[CLAassistant]

All committers have signed the CLA.

[NLPJCL]

Some issues should be held for now. An example of search-r1 should be provided under agentlloop instead of modifying the current one, and some experimental results should be provided.