keras-team/keras · error · ValueError

{name} must be >= 0. Received: {name}={value}

Error message

{name} must be >= 0. Received: {name}={value}

What it means

Shared validator for the pad/crop images ops: any explicitly supplied padding/cropping amount (or related numeric arg) must be non-negative. A negative value is rejected before any tensor op because negative padding/cropping is undefined.

Source

Thrown at keras/src/ops/image.py:1536

    """
    if any_symbolic_tensors((inputs, coordinates)):
        return MapCoordinates(
            order,
            fill_mode,
            fill_value,
        ).symbolic_call(inputs, coordinates)
    return backend.image.map_coordinates(
        inputs,
        coordinates,
        order,
        fill_mode,
        fill_value,
    )


def _validate_non_negative(value, name):
    if value is not None and value < 0:
        raise ValueError(f"{name} must be >= 0. Received: {name}={value}")


def _validate_pad_images_args(
    top_padding,
    left_padding,
    bottom_padding,
    right_padding,
    target_height,
    target_width,
):
    if [top_padding, bottom_padding, target_height].count(None) != 1:
        raise ValueError(
            "Must specify exactly two of "
            "top_padding, bottom_padding, target_height. "
            f"Received: top_padding={top_padding}, "
            f"bottom_padding={bottom_padding}, "
            f"target_height={target_height}"
        )

View on GitHub (pinned to 7a34a03db6)

Solutions

1. Use crop_images to shrink, pad_images to grow — never negative pad values
2. If deriving padding from target sizes, clamp: max(0, target - height - other) or switch to specifying target sizes directly
3. Validate all padding/cropping args against 0 before the call

Example fix

# before
pad_images(img, top_padding=height - target_height, ...)  # negative

# after
crop_images(img, top_cropping=target_height - height if False else 0, ...)  # or just:
pad_images(img, target_height=target_h, target_width=target_w, ...)  # derive non-negative pads

for name, v in [('top_padding', top_padding), ...]:
    if v is not None and v < 0:
        raise ValueError(f'{name} must be >= 0')

def all_non_negative(vals) -> bool:
    return all(v is None or v >= 0 for v in vals)

When it happens

Trigger: keras.ops.image.pad_images / crop_images (or the corresponding layers) with a negative top/bottom/left/right padding or cropping value, e.g. top_padding=-4, often computed as target - height - other_side and going negative.

Common situations: Using negative padding to emulate cropping; target_height smaller than the image so derived padding becomes negative; arithmetic on config values that can go below zero for some samples in a pipeline.

Understand the failure class

Background: "Must be a positive integer", "Invalid value", "Unsupported": the invalid-argument-value error family, when a library rejects the value you pass — this error's family across 35 libraries.

Related errors

• Must specify exactly two of top_padding, bottom_padding, tar
• Must specify exactly two of left_padding, right_padding, tar
• The `weights` argument should be either `None` (random initi
• `padding` should have two elements. Received: padding={paddi
• Invalid padding: {padding}

AI-assisted analysis of keras-team/keras@7a34a03db6 (2026-08-25). Data as JSON: /api/errors/c9b4acef74964d27. Report an issue: GitHub.