vision-base")

        >>> inputs = image_processor(images=image, return_tensors="pt")
        >>> outputs = model(**inputs)
        >>> # logits are of shape (batch_size, num_labels, height, width)
        >>> logits = outputs.logits
        ```NTrš