keras-team/keras · error · ValueError

Input should have its {axes} axes fully-defined. Received: i

Error message

Input should have its {axes} axes fully-defined. Received: input.shape = {real.shape}

What it means

fft2 needs both FFT axes (-2 and -1) fully defined at symbolic time; if either of the last two dimensions is None, the output shape cannot be computed and Keras raises. The message interpolates the axes tuple, e.g. 'Input should have its (-2, -1) axes fully-defined'.

Source

Thrown at keras/src/ops/math.py:595

        if real.shape != imag.shape:
            raise ValueError(
                "Input `x` should be a tuple of two tensors - real and "
                "imaginary. Both the real and imaginary parts should have the "
                f"same shape. Received: x[0].shape = {real.shape}, "
                f"x[1].shape = {imag.shape}"
            )
        # We are calculating 2D FFT. Hence, rank >= 2.
        if len(real.shape) < 2:
            raise ValueError(
                f"Input should have rank >= 2. "
                f"Received: input.shape = {real.shape}"
            )

        # The axes along which we are calculating FFT should be fully-defined.
        m = real.shape[axes[0]]
        n = real.shape[axes[1]]
        if m is None or n is None:
            raise ValueError(
                f"Input should have its {axes} axes fully-defined. "
                f"Received: input.shape = {real.shape}"
            )

        return (
            KerasTensor(shape=real.shape, dtype=real.dtype),
            KerasTensor(shape=imag.shape, dtype=imag.dtype),
        )

    def call(self, x):
        return backend.math.fft2(x)


@keras_export("keras.ops.fft2")
def fft2(x):
    """Computes the 2D Fast Fourier Transform along the last two axes of input.

    Args:

View on GitHub (pinned to 7a34a03db6)

Solutions

1. Fix both spatial dims: keras.Input(shape=(256, 256))
2. Resize all images to a fixed resolution before the FFT (resizing layer or preprocessing)
3. Bucket variable-size inputs to fixed sizes

Example fix

# before
inputs = keras.Input(shape=(None, None, 1))
out = keras.ops.fft2((inputs[..., 0], inputs[..., 0]))
# after
inputs = keras.Input(shape=(256, 256, 1))
out = keras.ops.fft2((inputs[..., 0], inputs[..., 0]))

assert real.shape[-1] is not None and real.shape[-2] is not None

When it happens

Trigger: keras.Input(shape=(None, None)) feeding fft2 in a functional model; variable-height image inputs (dynamic H and W).

Common situations: Variable-resolution image models; eager code moved into a functional graph where dims become None.

Related errors

• Input should have its last dimension fully-defined. Received
• The `weights` argument should be either `None` (random initi
• RandomCrop requires the input to have a fully defined height
• All `axis` values to be kept must have a known shape. Receiv
• R2Score expects 2D inputs with shape (batch_size, output_dim

AI-assisted analysis of keras-team/keras@7a34a03db6 (2026-08-25). Data as JSON: /api/errors/6a1555651df7458f. Report an issue: GitHub.