keras-team/keras · error · ValueError

input_tensor.shape[2] must equal input_shape[1]; Received…

Error message

input_tensor.shape[2] must equal input_shape[1]; Received `input_tensor.shape={input_tensor.shape}`, `input_shape={input_shape}`

What it means

MobileNetV2 builder in keras/src/applications/mobilenet_v2.py: when both input_shape and input_tensor are supplied (channels_last layout), input_tensor.shape[2] (the spatial width) must equal input_shape[1]. The guard fires when the declared input shape and the actual shape of the provided input tensor disagree, which would silently build an inconsistent graph.

Solutions

1. Align input_shape with input_tensor.shape, e.g. input_shape=(224, 224, 3) for a (None, 224, 224, 3) tensor
2. Pass only one of input_shape or input_tensor and let the builder infer the rest
3. Recreate the input tensor via keras.Input(shape=input_shape) to guarantee consistency

When it happens

Trigger: Thrown at keras/src/applications/mobilenet_v2.py:155 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/1b690b2e07af0057. Report an issue: GitHub.

Appendix: source

Thrown at keras/src/applications/mobilenet_v2.py:155

            except ValueError:
                raise ValueError(
                    f"input_tensor: {input_tensor}"
                    "is not type input_tensor. "
                    f"Received `type(input_tensor)={type(input_tensor)}`"
                )
        if is_input_t_tensor:
            if backend.image_data_format() == "channels_first":
                if input_tensor.shape[1] != input_shape[1]:
                    raise ValueError(
                        "input_shape[1] must equal shape(input_tensor)[1] "
                        "when `image_data_format` is `channels_first`; "
                        "Received `input_tensor.shape="
                        f"{input_tensor.shape}`"
                        f", `input_shape={input_shape}`"
                    )
            else:
                if input_tensor.shape[2] != input_shape[1]:
                    raise ValueError(
                        "input_tensor.shape[2] must equal input_shape[1]; "
                        "Received `input_tensor.shape="
                        f"{input_tensor.shape}`, "
                        f"`input_shape={input_shape}`"
                    )
        else:
            raise ValueError(
                "input_tensor is not a Keras tensor; "
                f"Received `input_tensor={input_tensor}`"
            )

    # If input_shape is None, infer shape from input_tensor.
    if input_shape is None and input_tensor is not None:
        try:
            backend.is_keras_tensor(input_tensor)
        except ValueError:
            raise ValueError(
                "input_tensor must be a valid Keras tensor type; "

View on GitHub (pinned to 7a34a03db6)