error of prompt in vllm

[lhbjyaqi12]

hello,I run this eval.sh:

CUDA_VISIBLE_DEVICES=0,1 python scripts/eval_vllm.py
--model_path "lingcco/fakeVLM"
--val_batch_size 16
--workers 16
--output_path results/fakevlm.json
--test_json_file "/home/work/xueyunqi/fakevlm/FakeVLM-main/test.json"
--data_base_test "/home/work/xueyunqi/fakevlm/FakeVLM-main/test"

after downloading the model from huggingface, there exists error like this:

(EngineCore_DP0 pid=26617) RuntimeError: Expected there to be 1 prompt placeholders corresponding to 1 image items, but instead found 0 prompt placeholders! Make sure the implementation of _call_hf_processor and _get_mm_fields_config are consistent with each other.
(EngineCore_DP0 pid=26617) ERROR 01-07 16:00:16 [multiproc_executor.py:231] Worker proc VllmWorker-0 died unexpectedly, shutting down executor.
/root/miniconda3/lib/python3.13/multiprocessing/resource_tracker.py:324: UserWarning: resource_tracker: There appear to be 1 leaked shared_memory objects to clean up at shutdown: {'/psm_1120bbfb'}
warnings.warn(
Traceback (most recent call last):
File "/home/work/xueyunqi/FakeVLM-main/scripts/eval_vllm.py", line 211, in
main()
~~~~^^
File "/home/work/xueyunqi/FakeVLM-main/scripts/eval_vllm.py", line 196, in main
model = load_model(args)
File "/home/work/xueyunqi/FakeVLM-main/scripts/eval_vllm.py", line 70, in load_model
llm = LLM(model=args.model_path,
dtype="float16",
max_model_len=800,
tensor_parallel_size=torch.cuda.device_count()
)
File "/root/miniconda3/lib/python3.13/site-packages/vllm/entrypoints/llm.py", line 351, in init
self.llm_engine = LLMEngine.from_engine_args(
~~~~~~~~~~~~~~~~~~~~~~~~~~^
engine_args=engine_args, usage_context=UsageContext.LLM_CLASS
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/root/miniconda3/lib/python3.13/site-packages/vllm/v1/engine/llm_engine.py", line 183, in from_engine_args
return cls(
vllm_config=vllm_config,
...<4 lines>...
multiprocess_mode=enable_multiprocessing,
)
File "/root/miniconda3/lib/python3.13/site-packages/vllm/v1/engine/llm_engine.py", line 109, in init
self.engine_core = EngineCoreClient.make_client(
~~~~~~~~~~~~~~~~~~~~~~~~~~~~^
multiprocess_mode=multiprocess_mode,
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
...<3 lines>...
log_stats=self.log_stats,
^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/root/miniconda3/lib/python3.13/site-packages/vllm/v1/engine/core_client.py", line 93, in make_client
return SyncMPClient(vllm_config, executor_class, log_stats)
File "/root/miniconda3/lib/python3.13/site-packages/vllm/v1/engine/core_client.py", line 648, in init
super().init(
~~~~~~~~~~~~~~~~^
asyncio_mode=False,
^^^^^^^^^^^^^^^^^^^
...<2 lines>...
log_stats=log_stats,
^^^^^^^^^^^^^^^^^^^^
)
^
File "/root/miniconda3/lib/python3.13/site-packages/vllm/v1/engine/core_client.py", line 477, in init
with launch_core_engines(vllm_config, executor_class, log_stats) as (
~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/root/miniconda3/lib/python3.13/contextlib.py", line 148, in exit
next(self.gen)
~~~~^^^^^^^^^^
File "/root/miniconda3/lib/python3.13/site-packages/vllm/v1/engine/utils.py", line 903, in launch_core_engines
wait_for_engine_startup(
~~~~~~~~~~~~~~~~~~~~~~~^
handshake_socket,
^^^^^^^^^^^^^^^^^
...<5 lines>...
coordinator.proc if coordinator else None,
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/root/miniconda3/lib/python3.13/site-packages/vllm/v1/engine/utils.py", line 960, in wait_for_engine_startup
raise RuntimeError(
...<3 lines>...
)
RuntimeError: Engine core initialization failed. See root cause above. Failed core proc(s): {'EngineCore_DP0': 1}

Can you help explain and solve it, ty

[lingcco]

Hi, we were using version 0.7.2 of VLLM before. You can try switching to this version to see if the similar error persists.