jax-ml/jax · error · TypeError

Cannot interpret '{dtype}' as a data type.\n\nDid you accide

Error message

Cannot interpret '{dtype}' as a data type.\n\nDid you accidentally write `jax.numpy.empty({shape}, {dtype})` when you meant `jax.numpy.empty(({shape}, {dtype}))`, i.e. with a single tuple argument for the shape?

What it means

jnp.empty's guard for the single-tuple typo: jnp.empty((2, 3, jnp.float32)) packs dtype into the shape tuple, causing dtype interpretation to fail; JAX raises with a hint suggesting jnp.empty((shape, dtype)) be written as separate arguments — note empty's unfilled memory means the wrong call silently produces wrong shapes otherwise.

Source

Thrown at jax/_src/numpy/array_creation.py:188

    Array of the specified shape and dtype, with the given device/sharding if specified.

  See also:
    - :func:`jax.lax.empty`
    - :func:`jax.numpy.empty_like`
    - :func:`jax.numpy.zeros`
    - :func:`jax.numpy.ones`
    - :func:`jax.numpy.full`

  Examples:
    >>> jnp.empty(4)  # doctest: +SKIP
    Array([0., 0., 0., 0.], dtype=float32)
    >>> jnp.empty((2, 3), dtype=bool)  # doctest: +SKIP
    Array([[False, False, False],
           [False, False, False]], dtype=bool)

  .. _explicit sharding: https://docs.jax.dev/en/latest/parallel.html
  """
  if (m := _check_forgot_shape_tuple("empty", shape, dtype)): raise TypeError(m)
  dtype = dtypes.check_and_canonicalize_user_dtype(
      float if dtype is None else dtype, "empty")
  shape = canonicalize_shape(shape)
  if device is not None and out_sharding is None:
    # lax.empty does not accept SingleDeviceSharding, so we use api.device_put.
    return api.device_put(lax.empty(shape, dtype), device)
  else:
    final_sharding = util.choose_device_or_out_sharding(
        device, out_sharding, 'jnp.empty')
    return lax.empty(shape, dtype, out_sharding=final_sharding)


def _check_forgot_shape_tuple(name, shape, dtype) -> str | None:
  if isinstance(dtype, int) and isinstance(shape, int):
    return (f"Cannot interpret '{dtype}' as a data type."
            f"\n\nDid you accidentally write "
            f"`jax.numpy.{name}({shape}, {dtype})` "
            f"when you meant `jax.numpy.{name}(({shape}, {dtype}))`, i.e. "

View on GitHub (pinned to 1e1c6a8fc0)

Solutions

Split into two arguments: jnp.empty((2, 3), jnp.float32)

Example fix

# before
a = jnp.empty((2, 3, jnp.float32))
# after
a = jnp.empty((2, 3), jnp.float32)

Defensive patterns

Strategy: validation

Validate before calling

assert all(isinstance(d, int) for d in shape), 'shape tuple should contain only ints'

Type guard

null

Try / catch

null

Prevention

Split shape and dtype into separate arguments
Check parenthesis placement when editing one-line creation calls

When it happens

Trigger: jnp.empty((2, 3, dtype)) — a single tuple containing shape dims plus a dtype element.

Common situations: One-line shape/dtype edits with misplaced parentheses; templated array creation code shared across zeros/ones/empty.

Related errors

AI-assisted analysis of jax-ml/jax@1e1c6a8fc0 (2026-08-27). Data as JSON: /api/errors/7547f5c6b5809baf. Report an issue: GitHub.