PaddlePaddle/PaddleOCR · error · TypeError

The input data is inconsistent with expectations.

Error message

The input data is inconsistent with expectations.

What it means

Raised by the ocr_cls hubserving module's predict method when the input does not match the expected shape. The module takes either `images` (list of HxWxC numpy arrays) or `paths` (list of image file path strings) — exactly one non-empty list. Any other combination or type falls through to this TypeError.

Source

Thrown at deploy/hubserving/ocr_cls/module.py:112

            images.append(img)
        return images

    def predict(self, images=[], paths=[]):
        """
        Get the text angle in the predicted images.
        Args:
            images (list(numpy.ndarray)): images data, shape of each is [H, W, C]. If images not paths
            paths (list[str]): The paths of images. If paths not images
        Returns:
            res (list): The result of text detection box and save path of images.
        """

        if images != [] and isinstance(images, list) and paths == []:
            predicted_data = images
        elif images == [] and isinstance(paths, list) and paths != []:
            predicted_data = self.read_images(paths)
        else:
            raise TypeError("The input data is inconsistent with expectations.")

        assert (
            predicted_data != []
        ), "There is not any image to be predicted. Please check the input data."

        img_list = []
        for img in predicted_data:
            if img is None:
                continue
            img_list.append(img)

        rec_res_final = []
        try:
            img_list, cls_res, predict_time = self.text_classifier(img_list)
            for dno in range(len(cls_res)):
                angle, score = cls_res[dno]
                rec_res_final.append(
                    {

View on GitHub (pinned to 2661c7c0ef)

Solutions

  1. Send exactly one non-empty JSON array: {"images": [...]} or {"paths": ["img1.jpg", "img2.jpg"]}.
  2. Wrap single items in a list (paths=["single.jpg"], not paths="single.jpg").
  3. When calling the class in Python, use predict(images=[arr], paths=[]) or predict(images=[], paths=[...]).

Example fix

# before
res = mod.predict(images=[], paths="doc.png")  # string, not list -> TypeError

# after
res = mod.predict(images=[], paths=["doc.png"])
Defensive patterns

Strategy: validation

Validate before calling

def valid_cls_payload(data: dict) -> bool:
    images, paths = data.get("images", []), data.get("paths", [])
    one = lambda v: isinstance(v, list) and len(v) > 0
    return (one(images) and not one(paths)) or (one(paths) and not one(images))

Type guard

def is_images_arg(v) -> bool:
    return isinstance(v, list) and len(v) > 0 and all(im is not None for im in v)

def is_paths_arg(v) -> bool:
    return isinstance(v, list) and len(v) > 0 and all(isinstance(p, str) for p in v)

Try / catch

try:
    res = mod.predict(images=images, paths=paths)
except TypeError as e:
    if "inconsistent" in str(e):
        return {"error": "invalid payload", "detail": str(e)}, 400
    raise

Prevention

When it happens

Trigger: Calling predict with both `images` and `paths` non-empty, both empty, `paths` as a bare string instead of a list, or `images` as a single ndarray instead of a list of arrays.

Common situations: POSTing {"paths": "single.jpg"} (string, not array) to the served endpoint; sending both keys in one request; adapting a script from PaddleOCR's command-line tools where a single image object is passed directly.

Related errors

AI-assisted analysis of PaddlePaddle/PaddleOCR@2661c7c0ef (2026-08-14). Data as JSON: /api/errors/44e4c2c75699dae1. Report an issue: GitHub.