keras-team/keras · error · ValueError

`inputs` argument cannot be empty. Received: inputs=

Error message

`inputs` argument cannot be empty. Received:
inputs={inputs}
outputs={outputs}

What it means

Raised by Functional.__init__ (keras.ops.function) when the `inputs` argument is empty/None while creating a Function via keras.ops.Function(inputs, outputs). Both inputs and outputs must be non-empty tensor structures.

Solutions

1. Pass at least one real KerasTensor as inputs
2. Check that tree.flatten(inputs) is non-empty before constructing

Example fix

# before
fn = keras.ops.Function([], out)

# after
fn = keras.ops.Function([x], out)

assert tree.flatten(inputs), 'inputs must be non-empty'

When it happens

Trigger: keras.ops.Function(inputs=[], outputs=t) or inputs=()

Common situations: Building keras.ops.Function programmatically where the input list ends up empty (e.g. a disconnected trace)

Related errors

• Function was called with an invalid input structure…
• Output with path ` ` is not connected to `inputs
• `outputs` argument cannot be empty. Received: inputs=
• Array inputs to associative_scan must have the same first…
• Cannot infer argument `num` from shape

AI-assisted analysis of keras-team/keras@7a34a03db6 (2026-08-25). Data as JSON: /api/errors/652b1a1cbad0ac93. Report an issue: GitHub.

Appendix: source

Thrown at keras/src/ops/function.py:63

    def __init__(self, inputs, outputs, name=None):
        super().__init__(name=name)

        if backend() == "tensorflow":
            # Temporary work around for
            # https://github.com/keras-team/keras/issues/931
            # This stop tensorflow from wrapping tf.function output in a
            # _DictWrapper object.
            _self_setattr_tracking = getattr(
                self, "_self_setattr_tracking", True
            )
            self._self_setattr_tracking = False
        self._inputs_struct = tree.map_structure(lambda x: x, inputs)
        self._outputs_struct = tree.map_structure(lambda x: x, outputs)
        self._inputs = tree.flatten(inputs)
        self._outputs = tree.flatten(outputs)
        if not self._inputs:
            raise ValueError(
                "`inputs` argument cannot be empty. Received:\n"
                f"inputs={inputs}\n"
                f"outputs={outputs}"
            )
        if not self._outputs:
            raise ValueError(
                "`outputs` argument cannot be empty. Received:\n"
                f"inputs={inputs}\n"
                f"outputs={outputs}"
            )

        if backend() == "tensorflow":
            self._self_setattr_tracking = _self_setattr_tracking

        (nodes, nodes_by_depth, operations, operations_by_depth) = map_graph(
            self._inputs, self._outputs
        )
        self._nodes = nodes

View on GitHub (pinned to 7a34a03db6)