for more detail.

        Returns:

        Example:

        ```python
        >>> from transformers import (
        ...     MgpstrProcessor,
        ...     MgpstrForSceneTextRecognition,
        ... )
        >>> import requests
        >>> from PIL import Image

        >>> # load image from the IIIT-5k dataset
        >>> url = "https://i.postimg.cc/ZKwLg2Gw/367-14.png"
        >>> image = Image.open(requests.get(url, stream=True).raw).convert("RGB")

        >>> processor = MgpstrProcessor.from_pretrained("alibaba-damo/mgp-str-base")
        >>> pixel_values = processor(images=image, return_tensors="pt").pixel_values

        >>> model = MgpstrForSceneTextRecognition.from_pretrained("alibaba-damo/mgp-str-base")

        >>> # inference
        >>> outputs = model(pixel_values)
        >>> out_strs = processor.batch_decode(outputs.logits)
        >>> out_strs["generated_text"]
        '["ticket"]'
        ```NrŪ