PaddlePredictorOption.__init__() takes 1 positional argument but 2 were given

[cnitker]

🔎 Search before asking

• I have searched the PaddleOCR Docs and found no similar bug report.
• I have searched the PaddleOCR Issues and found no similar bug report.
• I have searched the PaddleOCR Discussions and found no similar bug report.

🐛 Bug (问题描述)

运行官方提供的命令行：paddleocr ocr -i https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/general_ocr_002.png --use_doc_orientation_classify False --use_doc_unwarping False --use_textline_orientation False

提示以下信息：OCR版本 3.1.1 3.0.0 都试过都是报以下信息。运行环境Ubuntu22.04 python=3.10，此配置在window环境没问题

Traceback (most recent call last):
File "/home/jd001/miniconda3/envs/paddle-ocr/bin/paddleocr", line 8, in
sys.exit(console_entry())
File "/home/jd001/miniconda3/envs/paddle-ocr/lib/python3.10/site-packages/paddleocr/main.py", line 26, in console_entry
main()
File "/home/jd001/miniconda3/envs/paddle-ocr/lib/python3.10/site-packages/paddleocr/_cli.py", line 126, in main
_execute(args)
File "/home/jd001/miniconda3/envs/paddle-ocr/lib/python3.10/site-packages/paddleocr/_cli.py", line 115, in _execute
args.executor(args)
File "/home/jd001/miniconda3/envs/paddle-ocr/lib/python3.10/site-packages/paddleocr/_pipelines/ocr.py", line 630, in execute_with_args
perform_simple_inference(PaddleOCR, params)
File "/home/jd001/miniconda3/envs/paddle-ocr/lib/python3.10/site-packages/paddleocr/_utils/cli.py", line 62, in perform_simple_inference
wrapper = wrapper_cls(**init_params)
File "/home/jd001/miniconda3/envs/paddle-ocr/lib/python3.10/site-packages/paddleocr/_pipelines/ocr.py", line 161, in init
super().init(**base_params)
File "/home/jd001/miniconda3/envs/paddle-ocr/lib/python3.10/site-packages/paddleocr/_pipelines/base.py", line 66, in init
self.paddlex_pipeline = self._create_paddlex_pipeline()
File "/home/jd001/miniconda3/envs/paddle-ocr/lib/python3.10/site-packages/paddleocr/_pipelines/base.py", line 99, in _create_paddlex_pipeline
kwargs = prepare_common_init_args(None, self._common_args)
File "/home/jd001/miniconda3/envs/paddle-ocr/lib/python3.10/site-packages/paddleocr/_common_args.py", line 75, in prepare_common_init_args
pp_option = PaddlePredictorOption(
TypeError: PaddlePredictorOption.init() takes 1 positional argument but 2 were given

运行官方demo 提供

🏃‍♂️ Environment (运行环境)

OS : Ubuntu 22.04
python : 3.10

nvidia-cublas-cu12 12.6.4.1
nvidia-cuda-cccl-cu12 12.6.77
nvidia-cuda-cupti-cu12 12.6.80
nvidia-cuda-nvrtc-cu12 12.6.77
nvidia-cuda-runtime-cu12 12.6.77
nvidia-cudnn-cu12 9.5.1.17
nvidia-cufft-cu12 11.3.0.4
nvidia-cufile-cu12 1.11.1.6
nvidia-curand-cu12 10.3.7.77
nvidia-cusolver-cu12 11.7.1.2
nvidia-cusparse-cu12 12.5.4.2
nvidia-cusparselt-cu12 0.6.3
nvidia-nccl-cu12 2.25.1
nvidia-nvjitlink-cu12 12.6.85
nvidia-nvtx-cu12 12.6.77
openai 1.100.2
opencv-contrib-python 4.10.0.84
openpyxl 3.1.5
opt-einsum 3.3.0
orjson 3.11.2
packaging 25.0
paddleocr 3.1.1
paddlepaddle-gpu 3.0.0
paddlex 3.2.0
pandas 2.3.1
pillow 11.2.1

🌰 Minimal Reproducible Example (最小可复现问题的Demo)

paddleocr ocr -i https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/general_ocr_002.png --use_doc_orientation_classify False --use_doc_unwarping False --use_textline_orientation False

[ShulicTian]

解决了吗？我也碰到了

[SuiyueYoung]

一个解决方法就是 paddleocr 下 _common_args.py 里边第 75 行 PaddlePredictorOption() 内把 model_name 删了就行。PaddlePredictorOption 是通过 kwargs 来传参的，但是不知道要给哪个参数传递 model_name。 在测试用例里 model_name 是 None，不影响

def prepare_common_init_args(model_name, common_args):
    device = common_args["device"]
    if device is None:
        device = get_default_device()
    device_type, device_ids = parse_device(device)
    if device_ids is not None:
        device_id = device_ids[0]
    else:
        device_id = None

    init_kwargs = {}
    init_kwargs["use_hpip"] = common_args["enable_hpi"]
    init_kwargs["hpi_config"] = {
        "device_type": device_type,
        "device_id": device_id,
    }

    # model_name is deleted
    pp_option = PaddlePredictorOption(
        device_type=device_type, device_id=device_id
    )
    if device_type == "gpu":
        if common_args["use_pptrt"]:
            if common_args["pptrt_precision"] == "fp32":
                pp_option.run_mode = "trt_fp32"
            else:
                assert common_args["pptrt_precision"] == "fp16", common_args[
                    "pptrt_precision"
                ]
                pp_option.run_mode = "trt_fp16"
        else:
            pp_option.run_mode = "paddle"
    elif device_type == "cpu":
        enable_mkldnn = common_args["enable_mkldnn"]
        if enable_mkldnn:
            pp_option.run_mode = "mkldnn"
            pp_option.mkldnn_cache_capacity = common_args["mkldnn_cache_capacity"]
        else:
            pp_option.run_mode = "paddle"
        pp_option.cpu_threads = common_args["cpu_threads"]
    else:
        pp_option.run_mode = "paddle"
    init_kwargs["pp_option"] = pp_option

    return init_kwargs

[ShulicTian]

一个解决方法就是 paddleocr 下 _common_args.py 里边第 75 行 PaddlePredictorOption() 内把 model_name 删了就行。PaddlePredictorOption 是通过 kwargs 来传参的，但是不知道要给哪个参数传递 model_name。 在测试用例里 model_name 是 None，不影响

def prepare_common_init_args(model_name, common_args):
    device = common_args["device"]
    if device is None:
        device = get_default_device()
    device_type, device_ids = parse_device(device)
    if device_ids is not None:
        device_id = device_ids[0]
    else:
        device_id = None

    init_kwargs = {}
    init_kwargs["use_hpip"] = common_args["enable_hpi"]
    init_kwargs["hpi_config"] = {
        "device_type": device_type,
        "device_id": device_id,
    }

    # model_name is deleted
    pp_option = PaddlePredictorOption(
        device_type=device_type, device_id=device_id
    )
    if device_type == "gpu":
        if common_args["use_pptrt"]:
            if common_args["pptrt_precision"] == "fp32":
                pp_option.run_mode = "trt_fp32"
            else:
                assert common_args["pptrt_precision"] == "fp16", common_args[
                    "pptrt_precision"
                ]
                pp_option.run_mode = "trt_fp16"
        else:
            pp_option.run_mode = "paddle"
    elif device_type == "cpu":
        enable_mkldnn = common_args["enable_mkldnn"]
        if enable_mkldnn:
            pp_option.run_mode = "mkldnn"
            pp_option.mkldnn_cache_capacity = common_args["mkldnn_cache_capacity"]
        else:
            pp_option.run_mode = "paddle"
        pp_option.cpu_threads = common_args["cpu_threads"]
    else:
        pp_option.run_mode = "paddle"
    init_kwargs["pp_option"] = pp_option

    return init_kwargs

改源码我倒是知道，debug一下就看到了，主要是官方的案例为啥会失效

[ShulicTian]

我大概知道啥问题了，paddlex默认装了3.2.0的，要和安装版本3.1.1一致，靠

[ShulicTian]

果然，安装3.1.1的paddlex就带model_name这个参数了

[SuiyueYoung]

对，现在 paddleOCR 自动安装 paddlex 是 3.2 版本，把他改回 3.1 版本就好了

[WeizhouXueBenjamin]

我全部回退到了3.1.0， 还是一样报错啊 PaddlePredictorOption.__init__() takes 1 positional argument but 2 were given

[SuiyueYoung]

paddlex 3.1.0 的 paddlex.inference.utils.pp_option 下 PaddlePredictorOption 的构造方法如下：

def __init__(self, model_name=None, **kwargs):
        super().__init__()
        self._is_default_run_mode = True
        self._model_name = model_name
        self._cfg = {}
        self._init_option(**kwargs)
        self._changed = False

采用这个方法绝对不会错的。

[pfeak]

我回退版本后如上可以运行。

[Bobholamovic]

抱歉，这是因为PaddleX 3.2引入了一个接口变动，而PaddleOCR 3.1的元信息中没有对依赖的PaddleX版本上界做出限制。当前请考虑在安装时指定PaddleX的版本<3.2，例如pip install paddleocr "paddlex<3.2"。我们将在近期发布PaddleOCR 3.2（该版本可以适配PaddleX 3.2）并完善Python包的元信息。

[Shashank15newcold]

downgrading paddlex to 3.1.1, ensuring same version match with paddlepaddle-gpu and paddleocr might solve the issue. It solved for me.

[alichohann]

Any reliable solution for this problem?

[Shashank15newcold]

In 3.2.0, the constructor for PaddlePredictorOption changed — it no longer accepts two arguments.
So this solution is reliable, and it should solve the problem