keras-team/keras · error · ValueError

`inputs` should be a list. Received: inputs={inputs}

Error message

`inputs` should be a list. Received: inputs={inputs}

What it means

Concatenate.compute_mask validates that inputs is a list/tuple when a mask is supplied. A single tensor input (or non-sequence) together with a mask argument triggers this error.

Source

Thrown at keras/src/layers/merging/concatenate.py:128

                f"Received: input_shape={input_shape}"
            )
        input_shapes = input_shape
        output_shape = list(input_shapes[0])

        for shape in input_shapes[1:]:
            if output_shape[self.axis] is None or shape[self.axis] is None:
                output_shape[self.axis] = None
                break
            output_shape[self.axis] += shape[self.axis]
        return tuple(output_shape)

    def compute_mask(self, inputs, mask=None):
        if mask is None:
            return None
        if not isinstance(mask, (tuple, list)):
            raise ValueError(f"`mask` should be a list. Received mask={mask}")
        if not isinstance(inputs, (tuple, list)):
            raise ValueError(
                f"`inputs` should be a list. Received: inputs={inputs}"
            )
        if len(mask) != len(inputs):
            raise ValueError(
                "The lists `inputs` and `mask` should have the same length. "
                f"Received: inputs={inputs} of length {len(inputs)}, and "
                f"mask={mask} of length {len(mask)}"
            )
        if all(m is None for m in mask):
            return None
        # Make a list of masks while making sure
        # the dimensionality of each mask
        # is the same as the corresponding input.
        masks = []
        for input_i, mask_i in zip(inputs, mask):
            if mask_i is None:
                # Input is unmasked. Append all 1s to masks,
                masks.append(ops.ones_like(input_i, dtype="bool"))

View on GitHub (pinned to 7a34a03db6)

Solutions

1. Always pass inputs as a list matching the layer's call structure
2. Mirror the exact input structure used in __call__ when calling compute_mask
3. Add a structure check in wrappers before delegating

Example fix

# before
m = concat.compute_mask(x, mask=[m1, m2])

# after
m = concat.compute_mask([x1, x2], mask=[m1, m2])

assert isinstance(inputs, (list, tuple)), 'inputs must be a list when mask is given'

def is_input_list(x) -> bool:
    return isinstance(x, (list, tuple))

When it happens

Trigger: compute_mask(x, mask=[m1, m2]) called with a bare tensor; custom subclasses forwarding unwrapped inputs to the parent implementation.

Common situations: Direct compute_mask calls in tests or shape utilities; custom merge subclasses with altered input arity.

Related errors

• `inputs` should be a list. Received: inputs={inputs}
• A `Concatenate` layer should be called on a list of at least
• A `Concatenate` layer should be called on a list of inputs.
• `mask` should be a list. Received mask={mask}
• The lists `inputs` and `mask` should have the same length. R

AI-assisted analysis of keras-team/keras@7a34a03db6 (2026-08-25). Data as JSON: /api/errors/7bf4f18924714c24. Report an issue: GitHub.