[HELP]推理奖励模型报错，感谢大家，求教qwen基座rm后的模型如何vllm推理

[EvilCalf]

Exception in callback VllmEngine.patch_remove_log.<locals>.new_log_task_completion(error_callback=<bound method...7efd7bbd70b0>>)(<Task finishe...dimensions'")>) at /root/paddlejob/workspace/env_run/ms-swift/swift/llm/infer/infer_engine/vllm_engine.py:497
handle: <Handle VllmEngine.patch_remove_log.<locals>.new_log_task_completion(error_callback=<bound method...7efd7bbd70b0>>)(<Task finishe...dimensions'")>) at /root/paddlejob/workspace/env_run/ms-swift/swift/llm/infer/infer_engine/vllm_engine.py:497>
Traceback (most recent call last):
  File "/root/miniconda3/envs/swift/lib/python3.12/asyncio/events.py", line 84, in _run
    self._context.run(self._callback, *self._args)
  File "/root/paddlejob/workspace/env_run/ms-swift/swift/llm/infer/infer_engine/vllm_engine.py", line 499, in new_log_task_completion
    return_value = task.result()
                   ^^^^^^^^^^^^^
  File "/root/miniconda3/envs/swift/lib/python3.12/site-packages/vllm/engine/async_llm_engine.py", line 863, in run_engine_loop
    result = task.result()
             ^^^^^^^^^^^^^
  File "/root/miniconda3/envs/swift/lib/python3.12/site-packages/vllm/engine/async_llm_engine.py", line 786, in engine_step
    request_outputs = await self.engine.step_async(virtual_engine)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/swift/lib/python3.12/site-packages/vllm/engine/async_llm_engine.py", line 356, in step_async
    outputs = await self.model_executor.execute_model_async(
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/swift/lib/python3.12/site-packages/vllm/executor/executor_base.py", line 265, in execute_model_async
    output = await make_async(self.execute_model)(execute_model_req)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/swift/lib/python3.12/concurrent/futures/thread.py", line 58, in run
    result = self.fn(*self.args, **self.kwargs)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/swift/lib/python3.12/site-packages/vllm/executor/executor_base.py", line 140, in execute_model
    output = self.collective_rpc("execute_model",
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/swift/lib/python3.12/site-packages/vllm/executor/uniproc_executor.py", line 56, in collective_rpc
    answer = run_method(self.driver_worker, method, args, kwargs)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/swift/lib/python3.12/site-packages/vllm/utils.py", line 2456, in run_method
    return func(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/swift/lib/python3.12/site-packages/vllm/worker/worker_base.py", line 420, in execute_model
    output = self.model_runner.execute_model(
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/swift/lib/python3.12/site-packages/torch/utils/_contextlib.py", line 116, in decorate_context
    return func(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/swift/lib/python3.12/site-packages/vllm/worker/pooling_model_runner.py", line 148, in execute_model
    self.model.pooler(hidden_states=hidden_or_intermediate_states,
  File "/root/miniconda3/envs/swift/lib/python3.12/site-packages/vllm/model_executor/models/adapters.py", line 80, in pooler
    return self._pooler(hidden_states, pooling_metadata)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/swift/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1739, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/swift/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1750, in _call_impl
    return forward_call(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/swift/lib/python3.12/site-packages/vllm/model_executor/layers/pooler.py", line 100, in forward
    pooled_data = self.head(pooled_data, pooling_metadata)
                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/swift/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1739, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/swift/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1750, in _call_impl
    return forward_call(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/swift/lib/python3.12/site-packages/vllm/model_executor/layers/pooler.py", line 224, in forward
    pooling_param.dimensions
AttributeError: 'NoneType' object has no attribute 'dimensions'

vllm=0.8.5 ms-swift=3.4.0

CUDA_VISIBLE_DEVICES=1,2 swift infer \
    --model /root/paddlejob/workspace/env_run/xujiading/model_science/rm_2/v1-20250428-193329/checkpoint-937 \
    --infer_backend vllm \
    --dataset /root/paddlejob/workspace/env_run/xujiading/model_science/rm_2/v1-20250428-193329/val_dataset.jsonl \

[EvilCalf]

推理后端用pt是可以的

[EvilCalf]

用python的话出来的结果是生成式的内容，求教

[EvilCalf]

求大佬们以qwen为基座训出来的rm，Qwen2ForSequenceClassification/Qwen3ForSequenceClassification，怎么vllm推理啊