keras-team/keras · error · ValueError

When creating a Variable from an initializer, the `shape` ar

Error message

When creating a Variable from an initializer, the `shape` argument should be specified. Received: initializer={initializer} and shape={shape}

What it means

Error "When creating a Variable from an initializer, the `shape` argument should be specified. Received: initializer={initializer} and shape={shape}" thrown in keras-team/keras.

Source

Thrown at keras/src/backend/common/variables.py:170

        self._regularizer = None
        self._constraint = None
        self._trainable = bool(trainable)
        self._autocast = bool(autocast)
        self._aggregation = aggregation
        self._synchronization = synchronization
        self._layout = layout

        # `self._overwrite_with_gradient` is an internal property to determine
        # whether this variable should be overwritten by the computed gradient.
        # Ref: https://github.com/google/flax/blob/main/flax/linen/fp8_ops.py
        self._overwrite_with_gradient = False
        if isinstance(initializer, str):
            from keras.src import initializers

            initializer = initializers.get(initializer)
        if callable(initializer):
            if shape is None:
                raise ValueError(
                    "When creating a Variable from an initializer, "
                    "the `shape` argument should be specified. "
                    f"Received: initializer={initializer} "
                    f"and shape={shape}"
                )
        else:
            initializer = self._convert_to_tensor(initializer, dtype=dtype)
            # If dtype is None and `initializer` is an array, use its dtype.
            if dtype is None:
                dtype = initializer.dtype
        self._dtype = standardize_dtype(dtype)

        if in_stateless_scope():
            if callable(initializer):
                self._value = None
                self._initializer = initializer
                self._shape = self._validate_shape(shape)
                register_uninitialized_variable(self)

View on GitHub (pinned to 7a34a03db6)

When it happens

Trigger: Thrown at keras/src/backend/common/variables.py:170 when the library encounters an invalid state.

Common situations: See trigger scenarios.

AI-assisted analysis of keras-team/keras@7a34a03db6 (2026-08-25). Data as JSON: /api/errors/368d155cd5991caf. Report an issue: GitHub.