[Bug] [Module Name] vLLM Start Qwen2.5-32B-Instruct Error

[xydz]

Search before asking

• [X] I had searched in the issues and found no similar issues.

Operating system information

Linux

Python version information

3.10

DB-GPT version

main

Related scenes

• [ ] Chat Data
• [ ] Chat Excel
• [ ] Chat DB
• [ ] Chat Knowledge
• [ ] Model Management
• [ ] Dashboard
• [ ] Plugins

Installation Information

• [ ] Installation From Source

• [X] Docker Installation

• [ ] Docker Compose Installation

• [X] Cluster Installation

• [ ] AutoDL Image

• [ ] Other

Device information

GPU A800

Models information

LLM: Qwen2.5-32B-Instruct

What happened

Traceback (most recent call last): File "/usr/local/bin/dbgpt", line 8, in sys.exit(main()) File "/app/dbgpt/cli/cli_scripts.py", line 219, in main return cli() File "/usr/local/lib/python3.10/site-packages/click/core.py", line 1157, in call return self.main(*args, **kwargs) File "/usr/local/lib/python3.10/site-packages/click/core.py", line 1078, in main rv = self.invoke(ctx) File "/usr/local/lib/python3.10/site-packages/click/core.py", line 1688, in invoke return _process_result(sub_ctx.command.invoke(sub_ctx)) File "/usr/local/lib/python3.10/site-packages/click/core.py", line 1686, in invoke sub_ctx = cmd.make_context(cmd_name, args, parent=ctx) File "/usr/local/lib/python3.10/site-packages/click/core.py", line 943, in make_context self.parse_args(ctx, args) File "/usr/local/lib/python3.10/site-packages/click/core.py", line 1404, in parse_args parser = self.make_parser(ctx) File "/usr/local/lib/python3.10/site-packages/click/core.py", line 1315, in make_parser for param in self.get_params(ctx): File "/app/dbgpt/util/parameter_utils.py", line 866, in get_params dynamic_params = EnvArgumentParser._create_raw_click_option( File "/app/dbgpt/util/parameter_utils.py", line 464, in _create_raw_click_option combined_fields = _merge_dataclass_types( File "/app/dbgpt/util/parameter_utils.py", line 588, in _merge_dataclass_types _types = _dynamic_factory() File "/app/dbgpt/model/cli.py", line 413, in _model_dynamic_factory param_class = _dynamic_model_parser() File "/app/dbgpt/model/adapter/model_adapter.py", line 159, in _dynamic_model_parser param_class = llm_adapter.model_param_class() File "/app/dbgpt/model/adapter/vllm_adapter.py", line 58, in model_param_class return _build_parameter_class(descs) File "/app/dbgpt/util/parameter_utils.py", line 667, in _build_parameter_class result_class = dataclass(new_class) # type: ignore File "/usr/local/lib/python3.10/dataclasses.py", line 1185, in dataclass return wrap(cls) File "/usr/local/lib/python3.10/dataclasses.py", line 1176, in wrap return _process_class(cls, init, repr, eq, order, unsafe_hash, File "/usr/local/lib/python3.10/dataclasses.py", line 956, in _process_class cls_fields.append(_get_field(cls, name, type, kw_only)) File "/usr/local/lib/python3.10/dataclasses.py", line 813, in _get_field raise ValueError(f'mutable default {type(f.default)} for field ' ValueError: mutable default <class 'list'> for field ignore_patterns is not allowed: use default_factory

What you expected to happen

def model_param_class(self, model_type: str = None) -> BaseModelParameters:
    import argparse

    from vllm.engine.arg_utils import AsyncEngineArgs

    parser = argparse.ArgumentParser()
    parser = AsyncEngineArgs.add_cli_args(parser)
    parser.add_argument("--model_name", type=str, help="model name")
    parser.add_argument(
        "--model_path",
        type=str,
        help="local model path of the huggingface model to use",
    )
    parser.add_argument("--model_type", type=str, help="model type")
    # parser.add_argument("--device", type=str, default=None, help="device")
    # TODO parse prompt templete from `model_name` and `model_path`
    parser.add_argument(
        "--prompt_template",
        type=str,
        default=None,
        help="Prompt template. If None, the prompt template is automatically determined from model path",
    )

    descs = _extract_parameter_details(
        parser,
        "dbgpt.model.parameter.VLLMModelParameters",
        skip_names=["model"],
        overwrite_default_values={"trust_remote_code": True},
    )
    return _build_parameter_class(descs)
    
    
    dbgpt.model.parameter.VLLMModelParameters not find

How to reproduce

#!/bin/bash dbgpt start worker
--model_type vllm
--model_name qwen2.5-32b-instruct
--model_path /app/models/Qwen2.5-32B-Instruct
--port 8005
--gpu_memory_utilization 0.92
--max_model_len 4096
--controller_addr http://10.28.92.172:8000

Additional context

source version 0.5.10

Are you willing to submit PR?

• [ ] Yes I am willing to submit a PR!