docling-project/docling · error · TypeError

Unsupported score value type: {type(value)!r}

Error message

Unsupported score value type: {type(value)!r}

What it means

Raised as TypeError by HfVisionModelMixin._as_float when the score value is neither a Python Real (int/float), a numpy ndarray, nor a torch Tensor. The converter only accepts those three shapes of scalar-like values.

Source

Thrown at docling/models/inference_engines/common/hf_vision_base.py:133

            return float(value)

        if isinstance(value, np.ndarray):
            if value.size != 1:
                raise TypeError(
                    f"Expected scalar-like ndarray with size 1, got shape={value.shape}"
                )
            return float(value.reshape(-1)[0])

        import torch

        if isinstance(value, torch.Tensor):
            if value.numel() != 1:
                raise TypeError(
                    f"Expected scalar-like tensor with one element, got shape={tuple(value.shape)}"
                )
            return float(value.item())

        raise TypeError(f"Unsupported score value type: {type(value)!r}")

    @staticmethod
    def _as_int(value: Any) -> int:
        if isinstance(value, Integral):
            return int(value)

        if isinstance(value, np.ndarray):
            if value.size != 1:
                raise TypeError(
                    f"Expected scalar-like ndarray with size 1, got shape={value.shape}"
                )
            return int(value.reshape(-1)[0])

        import torch

        if isinstance(value, torch.Tensor):
            if value.numel() != 1:
                raise TypeError(

View on GitHub (pinned to 61d76f1ff3)

Solutions

1. Convert to float before passing: float(value) for lists of length 1, or index element [0].
2. Convert third-party tensors to numpy first (value.numpy()).
3. Keep your post-processor emitting only Python scalars, np.ndarray size-1, or torch.Tensor numel-1 values.

Example fix

# before
conf = model._as_float([0.93])  # list -> TypeError

# after
conf = model._as_float(0.93)  # or float(scores[i])

import numbers
import numpy as np
ok = isinstance(value, numbers.Real) or isinstance(value, np.ndarray) or _is_torch_tensor(value)
assert ok, f'unsupported score type {type(value)!r}'

import numbers
import numpy as np

def is_convertible_score(value) -> bool:
    if isinstance(value, (numbers.Real, np.ndarray)):
        return True
    try:
        import torch
        return torch.is_tensor(value)
    except ImportError:
        return False

When it happens

Trigger: Passing e.g. a Python list, dict, string, or a third-party array type (jax, tf.Tensor) as a score value.

Common situations: Swapping inference backends so outputs arrive as an unsupported container; test fixtures injecting plain lists as fake scores.

Related errors

• Expected scalar-like ndarray with size 1, got shape={value.s
• Expected scalar-like tensor with one element, got shape={tup
• Unsupported label value type: {type(value)!r}
• Unsupported input type: {type(self.path_or_stream)}
• Unexpected: {type(self.path_or_stream)=}

AI-assisted analysis of docling-project/docling@61d76f1ff3 (2026-08-14). Data as JSON: /api/errors/7d4e635edb5ac12d. Report an issue: GitHub.