keras-team/keras · error · ValueError

Cannot do batch_dot on inputs with rank < 2. Received inputs

Error message

Cannot do batch_dot on inputs with rank < 2. Received inputs with shapes {x_shape} and {y_shape}.

What it means

The batch_dot operation (used by Keras's Dot merge layer) requires both operands to have rank >= 2 (a batch dim plus at least one feature dim). This error fires in Dot._merge_function -> batch_dot when either input is a 1-D vector.

Source

Thrown at keras/src/layers/merging/dot.py:60

            integer. The sizes of `x.shape[axes[0]]` and `y.shape[axes[1]]`
            should be equal.
            Note that axis `0` (the batch axis) cannot be included.

    Returns:
        A tensor with shape equal to the concatenation of `x`'s shape
        (less the dimension that was summed over) and `y`'s shape (less the
        batch dimension and the dimension that was summed over). If the final
        rank is 1, we reshape it to `(batch_size, 1)`.
    """

    x_shape = x.shape
    y_shape = y.shape

    x_ndim = len(x_shape)
    y_ndim = len(y_shape)

    if x_ndim < 2 or y_ndim < 2:
        raise ValueError(
            f"Cannot do batch_dot on inputs "
            f"with rank < 2. "
            f"Received inputs with shapes "
            f"{x_shape} and {y_shape}."
        )

    x_batch_size = x_shape[0]
    y_batch_size = y_shape[0]

    if x_batch_size is not None and y_batch_size is not None:
        if x_batch_size != y_batch_size:
            raise ValueError(
                f"Cannot do batch_dot on inputs "
                f"with different batch sizes. "
                f"Received inputs with shapes "
                f"{x_shape} and {y_shape}."
            )
    if isinstance(axes, int):

View on GitHub (pinned to 7a34a03db6)

Solutions

1. Expand vectors to rank 2 before the Dot: x = keras.ops.expand_dims(x, -1)
2. Use keras.ops.sum(x * y, axis=-1) or a Dense(1) for vector inner products instead of Dot
3. Check upstream layers (squeeze, reduce) that might drop a dimension

Example fix

# before
out = layers.Dot(axes=-1)([x, y])  # x shape (None,) -> ValueError

# after
x = layers.Reshape((1,))(x)  # or expand_dims to rank 2
out = layers.Dot(axes=-1)([x, y])

assert len(x.shape) >= 2 and len(y.shape) >= 2, 'batch_dot needs rank >= 2'

def batch_dot_ok(x, y) -> bool:
    return len(x.shape) >= 2 and len(y.shape) >= 2

When it happens

Trigger: Calling layers.Dot(axes=1)([x, y]) where x or y is shape (None,) or (batch_size,); using batched products on plain vectors.

Common situations: Dotting raw embedding outputs without a time/channel axis; mixing up Dot with inner products on flattened vectors; inputs unexpectedly squeezed upstream.

Related errors

• A `Concatenate` layer requires inputs with matching shapes e
• Cannot do batch_dot on inputs with rank < 2. Received inputs
• Cannot do batch_dot on inputs with different batch sizes. Re
• Multiple target dimensions are not supported. Expected: None
• Cannot perform batch_dot over axis 0. If your inputs are not

AI-assisted analysis of keras-team/keras@7a34a03db6 (2026-08-25). Data as JSON: /api/errors/3a90328f0eb4f913. Report an issue: GitHub.