keras-team/keras · error · ValueError

The `weights` argument should be either `None` (random initi

Error message

The `weights` argument should be either `None` (random initialization), `imagenet` (pre-training on ImageNet), or the path to the weights file to be loaded.Received: weights={weights}

What it means

The cropping operation's target_height must be >= 0; a negative final cropped height is rejected during output-shape computation.

Source

Thrown at keras/src/applications/efficientnet_v2.py:896

            only to be specified if `include_top` is `True`, and if no `weights`
            argument is specified.
        classifier_activation: A string or callable. The activation function to
            use on the "top" layer. Ignored unless `include_top=True`. Set
            `classifier_activation=None` to return the logits of the "top"
            layer.
        include_preprocessing: Boolean, whether to include the preprocessing
            layer (`Rescaling`) at the bottom of the network.
            Defaults to `True`.

    Returns:
        A model instance.
    """

    if blocks_args == "default":
        blocks_args = DEFAULT_BLOCKS_ARGS[name]

    if not (weights in {"imagenet", None} or file_utils.exists(weights)):
        raise ValueError(
            "The `weights` argument should be either "
            "`None` (random initialization), `imagenet` "
            "(pre-training on ImageNet), "
            "or the path to the weights file to be loaded."
            f"Received: weights={weights}"
        )

    if weights == "imagenet" and include_top and classes != 1000:
        raise ValueError(
            'If using `weights="imagenet"` with `include_top`'
            " as true, `classes` should be 1000"
        )

    # Determine proper input shape
    input_shape = imagenet_utils.obtain_input_shape(
        input_shape,
        default_size=default_size,
        min_size=32,

View on GitHub (pinned to 7a34a03db6)

Solutions

1. Validate target_height >= 0 before calling
2. Fix the arithmetic that produced the negative value
3. Skip cropping for degenerate sizes

Example fix

# before
out = ops.image.crop_images(img, target_height=(h - 2 * c, w - 2 * c))  # h < 2c
# after
if h - 2 * c < 0:
    raise ValueError('image too small to crop')
out = ops.image.crop_images(img, target_height=(h - 2 * c, w - 2 * c))

target_height = int(target_height)
if target_height < 0:
    raise ValueError('target_height must be >= 0')

When it happens

Trigger: Passing a negative target_height, often from a subtraction like target = size - 2*crop that underflows.

Common situations: Computed crop targets on small images; config typos; -1 sentinel dynamic dims leaking into the call.

Related errors

• Expected mode to be one of `caffe`, `tf` or `torch`. Receive
• weights_path undefined
• The `weights` argument should be either `None` (random initi
• If using `weights="imagenet"` as true, `classes` should be 1
• The number of repeats in `EfficientNet` must be > 0. Receive

AI-assisted analysis of keras-team/keras@7a34a03db6 (2026-08-25). Data as JSON: /api/errors/3c6612d55a982892. Report an issue: GitHub.