keras-team/keras · error · ValueError

Arguments `target` and `output` must have the same shape. Re

Error message

Arguments `target` and `output` must have the same shape. Received: target.shape={target.shape}, output.shape={output.shape}

What it means

Error "Arguments `target` and `output` must have the same shape. Received: target.shape={target.shape}, output.shape={output.shape}" thrown in keras-team/keras.

Source

Thrown at keras/src/backend/jax/ops/nn.py:971

        values = jnp.greater_equal(result.data, 0).astype(dtype)
        return jax_sparse.BCOO(
            (values, result.indices),
            shape=result.shape,
            indices_sorted=True,
            unique_indices=True,
        )
    return jnp.max(
        one_hot(cast(x, "int32"), num_classes, axis=axis, dtype=dtype),
        axis=reduction_axis,
    )


def categorical_crossentropy(target, output, from_logits=False, axis=-1):
    target = jnp.array(target)
    output = jnp.array(output)

    if target.shape != output.shape:
        raise ValueError(
            "Arguments `target` and `output` must have the same shape. "
            "Received: "
            f"target.shape={target.shape}, output.shape={output.shape}"
        )
    if len(target.shape) < 1:
        raise ValueError(
            "Arguments `target` and `output` must be at least rank 1. "
            "Received: "
            f"target.shape={target.shape}, output.shape={output.shape}"
        )

    if from_logits:
        log_prob = jax.nn.log_softmax(output, axis=axis)
    else:
        output = output / jnp.sum(output, axis, keepdims=True)
        output = jnp.clip(output, backend.epsilon(), 1.0 - backend.epsilon())
        log_prob = jnp.log(output)
    return -jnp.sum(target * log_prob, axis=axis)

View on GitHub (pinned to 7a34a03db6)

When it happens

Trigger: Thrown at keras/src/backend/jax/ops/nn.py:971 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/22ce2debd32ec525. Report an issue: GitHub.