keras-team/keras · error · ValueError

When setting `include_top=True` and loading `imagenet`…

Error message

When setting `include_top=True` and loading `imagenet` weights, `input_shape` should be {default_shape}.  Received: input_shape={input_shape}

What it means

obtain_input_shape in keras/src/applications/imagenet_utils.py: when loading pretrained "imagenet" weights with include_top=True, the classifier weights fix the spatial size, so input_shape must exactly equal the model's default shape (e.g. (224,224,3)). Any other shape — including correct channel count but wrong width/height — is rejected because the dense top cannot be reshaped to it.

Solutions

1. Pass input_shape equal to the model's default_size, e.g. (224, 224, 3), or omit input_shape entirely to use the default
2. Set include_top=False if you need custom spatial dimensions and will add your own head
3. Use random initialization (weights=None) if a non-standard input size is required with a top

When it happens

Trigger: Thrown at keras/src/applications/imagenet_utils.py:352 when the library encounters an invalid state.

Common situations: See trigger scenarios.

AI-assisted analysis of keras-team/keras@7a34a03db6 (2026-08-25). Data as JSON: /api/errors/71a897cdd69d1f36. Report an issue: GitHub.

Appendix: source

Thrown at keras/src/applications/imagenet_utils.py:352

            default_shape = (input_shape[0], default_size, default_size)
        else:
            if input_shape[-1] not in {1, 3}:
                warnings.warn(
                    "This model usually expects 1 or 3 input channels. "
                    "However, it was passed an input_shape "
                    f"with {input_shape[-1]} input channels.",
                    stacklevel=2,
                )
            default_shape = (default_size, default_size, input_shape[-1])
    else:
        if data_format == "channels_first":
            default_shape = (3, default_size, default_size)
        else:
            default_shape = (default_size, default_size, 3)
    if weights == "imagenet" and require_flatten:
        if input_shape is not None:
            if input_shape != default_shape:
                raise ValueError(
                    "When setting `include_top=True` "
                    "and loading `imagenet` weights, "
                    f"`input_shape` should be {default_shape}.  "
                    f"Received: input_shape={input_shape}"
                )
        return default_shape
    if input_shape:
        if data_format == "channels_first":
            if input_shape is not None:
                if len(input_shape) != 3:
                    raise ValueError(
                        "`input_shape` must be a tuple of three integers."
                    )
                if input_shape[0] != 3 and weights == "imagenet":
                    raise ValueError(
                        "The input must have 3 channels; Received "
                        f"`input_shape={input_shape}`"
                    )

View on GitHub (pinned to 7a34a03db6)