keras-team/keras · error · ValueError

`height` and `width` must be set if `format='xyxy'`.

Error message

`height` and `width` must be set if `format='xyxy'`.

What it means

clip_to_image_size clips bounding boxes to the image boundaries. Only 'xyxy' and 'rel_xyxy' formats are supported (others raise NotImplementedError). In absolute 'xyxy' coordinates the function needs the pixel height and width to clip against, so passing height=None or width=None with that format raises this ValueError.

Source

Thrown at keras/src/layers/preprocessing/image_preprocessing/bounding_boxes/bounding_box.py:116

        if source.startswith("rel_") and target.startswith("rel_"):
            source = source.replace("rel_", "", 1)
            target = target.replace("rel_", "", 1)
        to_xyxy_converter = to_xyxy_converters[source]
        from_xyxy_converter = from_xyxy_converters[target]
        in_xyxy_boxes = to_xyxy_converter(boxes, height, width)
        return from_xyxy_converter(in_xyxy_boxes, height, width)

    def clip_to_image_size(
        self,
        bounding_boxes,
        height=None,
        width=None,
        bounding_box_format="xyxy",
    ):
        if bounding_box_format not in ("xyxy", "rel_xyxy"):
            raise NotImplementedError
        if bounding_box_format == "xyxy" and (height is None or width is None):
            raise ValueError(
                "`height` and `width` must be set if `format='xyxy'`."
            )

        ops = self.backend
        boxes = bounding_boxes["boxes"]
        labels = bounding_boxes.get("labels", None)
        if width is not None:
            width = ops.cast(width, boxes.dtype)
        if height is not None:
            height = ops.cast(height, boxes.dtype)

        if bounding_box_format == "xyxy":
            x1, y1, x2, y2 = ops.numpy.split(boxes, 4, axis=-1)
            x1 = ops.numpy.clip(x1, 0, width)
            y1 = ops.numpy.clip(y1, 0, height)
            x2 = ops.numpy.clip(x2, 0, width)
            y2 = ops.numpy.clip(y2, 0, height)
            boxes = ops.numpy.concatenate([x1, y1, x2, y2], axis=-1)

View on GitHub (pinned to 7a34a03db6)

Solutions

1. Pass integer height and width matching the image the boxes refer to, e.g. clip_to_image_size(bb, height=416, width=416, bounding_box_format="xyxy").
2. If sizes are unknown, use format='rel_xyxy' so no height/width is needed.

Example fix

# before
clipped = clip_to_image_size(boxes, bounding_box_format="xyxy")
# after
clipped = clip_to_image_size(boxes, bounding_box_format="xyxy", height=img.shape[1], width=img.shape[2])

if bounding_box_format == "xyxy":
    assert height is not None and width is not None, "xyxy requires height and width"

When it happens

Trigger: Calling clip_to_image_size(boxes, bounding_box_format="xyxy") without height/width, or with only one of them set.

Common situations: Pipeline refactor from relative to absolute coordinates where the image-size arguments were dropped; calling the utility on tensors whose spatial dims are unknown so the caller skipped sizes.

Understand the failure class

Background: "missing required argument" and "the following required arguments were not provided": what required-argument errors mean and how to fix them — this error's family across 20 libraries.

Related errors

• When using relative bounding box formats (e.g. `rel_yxyx`) t
• `variance` must be length 4, got {variance}
• `encoding_format` should be one of 'center_xywh' or 'center_
• `encoded_format` should be 'center_xywh' or 'center_yxhw', b
• compute_iou() expects boxes1 to be batched, or to be unbatch

AI-assisted analysis of keras-team/keras@7a34a03db6 (2026-08-25). Data as JSON: /api/errors/e95d75ebc4eb19e7. Report an issue: GitHub.