jax-ml/jax · error · NotImplementedError

The 'out' argument to jnp.choose is not supported.

Error message

The 'out' argument to jnp.choose is not supported.

What it means

JAX arrays are immutable, so jnp.choose cannot write its broadcast selection into an out array. The wrapper rejects any non-None out argument.

Solutions

  1. Remove out= and use the return value
  2. If out semantics are needed, emulate with out = out.at[...].set(jnp.choose(...))

Example fix

// before
jnp.choose(a, choices, out=out)
// after
out = jnp.choose(a, choices)
Defensive patterns

Strategy: type-guard

Validate before calling

assert out is None

Prevention

When it happens

Trigger: jnp.choose(a, choices, out=buf) — code using numpy's out parameter for choose.

Common situations: Ported numpy code that preallocates outputs for performance; shared utility functions that thread an out parameter through calls.

Related errors


AI-assisted analysis of jax-ml/jax@1e1c6a8fc0 (2026-08-27). Data as JSON: /api/errors/8c440724576acb53. Report an issue: GitHub.

Appendix: source

Thrown at jax/_src/numpy/lax_numpy.py:4963

    In the more general case, ``choices`` may be a sequence of array-like
    objects with any broadcast-compatible shapes.

    >>> choice_1 = jnp.array([1, 2, 3, 4])
    >>> choice_2 = 99
    >>> choice_3 = jnp.array([[10],
    ...                       [20],
    ...                       [30]])
    >>> a = jnp.array([[0, 1, 2, 0],
    ...                [1, 2, 0, 1],
    ...                [2, 0, 1, 2]])
    >>> jnp.choose(a, [choice_1, choice_2, choice_3], mode='wrap')
    Array([[ 1, 99, 10,  4],
           [99, 20,  3, 99],
           [30,  2, 99, 30]], dtype=int32)
  """
  if out is not None:
    raise NotImplementedError("The 'out' argument to jnp.choose is not supported.")
  a, *choices = util.ensure_arraylike_tuple('choose', (a, *choices))
  if not issubdtype(a.dtype, np.integer):
    raise ValueError("`a` array must be integer typed")
  N = len(choices)

  if mode == 'raise':
    arr: Array = core.concrete_or_error(asarray, a,
      "The error occurred because jnp.choose was jit-compiled"
      " with mode='raise'. Use mode='wrap' or mode='clip' instead.")
    if reductions.any((arr < 0) | (arr >= N)):
      raise ValueError("invalid entry in choice array")
  elif mode == 'wrap':
    arr = asarray(a) % N
  elif mode == 'clip':
    arr = clip(a, 0, N - 1)
  else:
    raise ValueError(f"mode={mode!r} not understood. Must be 'raise', 'wrap', or 'clip'")

View on GitHub (pinned to 1e1c6a8fc0)