0.2.23 运行 weclone-cli server 报错

[quqixun]

• weclone : 0.2.23
• llamafactory : 0.9.3.dev0
• transformers : 4.52.1

运行 weclone-cli webchat-demo 可以正常使用；
但是运行 weclone-cli server 报错：

[WeClone] I | 11:29:47 | Loading configuration from: ./settings.jsonc
INFO 06-16 11:29:50 [importing.py:53] Triton module has been replaced with a placeholder.
INFO 06-16 11:29:50 [__init__.py:239] Automatically detected platform cuda.
[WeClone] I | 11:29:51 | Loading configuration from: ./settings.jsonc
Traceback (most recent call last):
  File "/home/q/miniconda3/envs/weclone/bin/weclone-cli", line 8, in <module>
    sys.exit(cli())
  File "/home/q/miniconda3/envs/weclone/lib/python3.10/site-packages/click/core.py", line 1442, in __call__
    return self.main(*args, **kwargs)
  File "/home/q/miniconda3/envs/weclone/lib/python3.10/site-packages/click/core.py", line 1363, in main
    rv = self.invoke(ctx)
  File "/home/q/miniconda3/envs/weclone/lib/python3.10/site-packages/click/core.py", line 1830, in invoke
    return _process_result(sub_ctx.command.invoke(sub_ctx))
  File "/home/q/miniconda3/envs/weclone/lib/python3.10/site-packages/click/core.py", line 1226, in invoke
    return ctx.invoke(self.callback, **ctx.params)
  File "/home/q/miniconda3/envs/weclone/lib/python3.10/site-packages/click/core.py", line 794, in invoke
    return callback(*args, **kwargs)
  File "/home/q/Projects/Clone/WeClone/weclone/cli.py", line 33, in wrapper
    return func(*args, **kwargs)
  File "/home/q/Projects/Clone/WeClone/weclone/cli.py", line 51, in new_runtime_wrapper
    return original_cmd_func(*args, **kwargs)
  File "/home/q/Projects/Clone/WeClone/weclone/cli.py", line 121, in server
    server_main()
  File "/home/q/Projects/Clone/WeClone/weclone/server/api_service.py", line 12, in main
    chat_model = ChatModel(config)
  File "/home/q/miniconda3/envs/weclone/lib/python3.10/site-packages/llamafactory/chat/chat_model.py", line 51, in __init__
    model_args, data_args, finetuning_args, generating_args = get_infer_args(args)
  File "/home/q/miniconda3/envs/weclone/lib/python3.10/site-packages/llamafactory/hparams/parser.py", line 413, in get_infer_args
    model_args, data_args, finetuning_args, generating_args = _parse_infer_args(args)
  File "/home/q/miniconda3/envs/weclone/lib/python3.10/site-packages/llamafactory/hparams/parser.py", line 187, in _parse_infer_args
    return _parse_args(parser, args, allow_extra_keys=allow_extra_keys)
  File "/home/q/miniconda3/envs/weclone/lib/python3.10/site-packages/llamafactory/hparams/parser.py", line 82, in _parse_args
    (*parsed_args, unknown_args) = parser.parse_args_into_dataclasses(args=args, return_remaining_strings=True)
  File "/home/q/miniconda3/envs/weclone/lib/python3.10/site-packages/transformers/hf_argparser.py", line 350, in parse_args_into_dataclasses
    args = file_args + args if args is not None else file_args + sys.argv[1:]
TypeError: can only concatenate list (not "WCInferConfig") to list

[quqixun]

修改 weclone/server/api_service.py 可以运行 weclone-cli server，仅供参考。

import os

import uvicorn
from llamafactory.api.app import create_app
from llamafactory.chat import ChatModel

from weclone.utils.config import load_config


def main():
    config = load_config("api_service")

    # chat_model = ChatModel(config)

    # --------------------------------------------
    # 将 config 转换为命令行参数列表
    args = []
    for key, value in config.__dict__.items():
        if value is not None:
            args.extend([f"--{key}", str(value)])
    chat_model = ChatModel(args)
    # --------------------------------------------
    
    app = create_app(chat_model)
    print("Visit http://localhost:{}/docs for API document.".format(os.environ.get("API_PORT", 8005)))
    uvicorn.run(app, host="0.0.0.0", port=int(os.environ.get("API_PORT", 8005)), workers=1)


if __name__ == "__main__":
    main()