huggingface/transformers · error · ValueError

Unsupported input image type {image_type}

Error message

Unsupported input image type {image_type}

What it means

Error "Unsupported input image type {image_type}" thrown in huggingface/transformers.

Source

Thrown at src/transformers/image_processing_backends.py:127

        elif isinstance(image_url_or_urls, str):
            return load_image_as_tensor(image_url_or_urls)
        elif is_valid_image(image_url_or_urls):
            return image_url_or_urls
        else:
            raise TypeError(f"only a single or a list of entries is supported but got type={type(image_url_or_urls)}")

    def process_image(
        self,
        image: ImageInput,
        do_convert_rgb: bool | None = None,
        input_data_format: str | ChannelDimension | None = None,
        device: Optional["torch.device"] = None,
        **kwargs: Unpack[ImagesKwargs],
    ) -> "torch.Tensor":
        """Process a single image for torchvision backend."""
        image_type = get_image_type(image)
        if image_type not in [ImageType.PIL, ImageType.TORCH, ImageType.NUMPY]:
            raise ValueError(f"Unsupported input image type {image_type}")

        if do_convert_rgb:
            image = self.convert_to_rgb(image)

        if image_type == ImageType.PIL:
            image = tvF.pil_to_tensor(image)
        elif image_type == ImageType.NUMPY:
            image = torch.from_numpy(image).contiguous()

        if image.ndim == 2:
            image = image.unsqueeze(0)

        if input_data_format is None:
            input_data_format = infer_channel_dimension_format(image)

        if input_data_format == ChannelDimension.LAST:
            image = image.permute(2, 0, 1).contiguous()

View on GitHub (pinned to a597f97485)

Solutions

1. Pass images as PIL images, numpy arrays, or torch tensors.
2. Convert the input with `load_image` or `Image.open` first.

When it happens

Trigger: Raised in image processing backend dispatch when the input image type has no registered backend handler.

Common situations: Feeding an unsupported object (e.g. a raw bytes buffer or unknown array type) as an image to the processor.

AI-assisted analysis of huggingface/transformers@a597f97485 (2026-08-14). Data as JSON: /api/errors/bdf237e14b581e75. Report an issue: GitHub.