keras-team/keras · error · ValueError

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

Error message

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

What it means

Symmetric to the inputs check, outputs must flatten to a non-empty structure; a Function with no outputs is meaningless and rejected at construction.

Source

Thrown at keras/src/ops/function.py:69

            # 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
        self._nodes_by_depth = nodes_by_depth
        self._operations = operations
        self._operations_by_depth = operations_by_depth

        # Run through graph to check all outputs are connected to the inputs.
        def empty_op_outputs(op, *args, **kwargs):

View on GitHub (pinned to 7a34a03db6)

Solutions

1. Ensure the op graph produces at least one output tensor
2. Pass a concrete output tensor instead of an empty container

Example fix

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

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

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

When it happens

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

Common situations: Programmatic graph construction with bugs that drop the computed tensors

Related errors

• `inputs` argument cannot be empty. Received: inputs={inputs}
• Output with path `{path}` is not connected to `inputs`
• Function was called with an invalid input structure. Expecte
• Sequential model '{self.name}' has no defined outputs yet.
• Array inputs to associative_scan must have the same first di

AI-assisted analysis of keras-team/keras@7a34a03db6 (2026-08-25). Data as JSON: /api/errors/6d356c75a29f61fd. Report an issue: GitHub.