keras-team/keras · error · ValueError

Argument `n` should be a positive integer. Received: n={n}

Error message

Argument `n` should be a positive integer. Received: n={n}

What it means

RepeatVector.__init__ rejects n <= 0 after the int check: repeating a vector zero or a negative number of times would produce an empty or invalid axis, so Keras fails fast at construction time.

Source

Thrown at keras/src/layers/reshaping/repeat_vector.py:35

    Args:
        n: Integer, repetition factor.

    Input shape:
        2D tensor with shape `(batch_size, features)`.

    Output shape:
        3D tensor with shape `(batch_size, n, features)`.
    """

    def __init__(self, n, **kwargs):
        super().__init__(**kwargs)
        if not isinstance(n, int) or isinstance(n, bool):
            raise TypeError(
                f"Expected an integer value for `n`, got {type(n)}."
            )
        if n <= 0:
            raise ValueError(
                f"Argument `n` should be a positive integer. Received: n={n}"
            )
        self.n = n
        self.input_spec = InputSpec(ndim=2)

    def compute_output_shape(self, input_shape):
        return (input_shape[0], self.n, input_shape[1])

    def call(self, inputs):
        input_shape = ops.shape(inputs)
        reshaped = ops.reshape(inputs, (input_shape[0], 1, input_shape[1]))
        return ops.repeat(reshaped, self.n, axis=1)

    def get_config(self):
        config = {"n": self.n}
        base_config = super().get_config()
        return {**base_config, **config}

View on GitHub (pinned to 7a34a03db6)

Solutions

1. Ensure the value feeding n is >= 1; add a max(1, ...) guard only if a 1-repeat is acceptable semantics
2. Trace where n comes from — usually an upstream calculation returning 0 (empty list length, division rounding down)
3. Validate configs at load time before layer construction

Example fix

# before
n = len(seq) - 1  # 0 when len(seq)==1
layer = RepeatVector(n=n)

# after
n = max(len(seq), 1)
layer = RepeatVector(n=n)

def validated_n(n):
    if not (isinstance(n, int) and not isinstance(n, bool)):
        raise TypeError('n must be int')
    if n <= 0:
        raise ValueError(f'n must be >= 1, got {n}')
    return n

def is_positive_int(v) -> bool:
    return isinstance(v, int) and not isinstance(v, bool) and v > 0

When it happens

Trigger: RepeatVector(n=0) or RepeatVector(n=-3). Typically the value comes from a computation or config that accidentally evaluates to zero or negative.

Common situations: Hyperparameter search proposing 0; deriving n from a length or batch-size expression that can be 0 for empty inputs; default/placeholder config values that were never replaced.

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

• Expected an integer value for `n`, got {type(n)}.
• Argument `size` should be a positive integer. Received: size
• Unknown activation function '{activation}' cannot be seriali
• Could not interpret activation function identifier: {identif
• ConvNeXt does not support the `channels_first` image data fo

AI-assisted analysis of keras-team/keras@7a34a03db6 (2026-08-25). Data as JSON: /api/errors/397b511bb1ad7c0f. Report an issue: GitHub.