jax-ml/jax · warning · ValueError

Unrecognized {kind=} expected one of {list(_dtype_kinds.keys

Error message

Unrecognized {kind=} expected one of {list(_dtype_kinds.keys())}, or a compatible input for jnp.dtype()

What it means

jax.dtypes.isdtype implements the Array API: kind may be a recognized kind string ('bool','signed integer',...), a dtype-like, or tuples thereof. A string that is neither a known kind nor constructible by np.dtype raises ValueError; non-string non-dtype inputs raise TypeError.

Source

Thrown at jax/_src/dtypes.py:672

      If ``kind`` is a tuple, then return True if dtype matches any entry of the tuple.

  Returns:
    True or False
  """
  the_dtype = np.dtype(dtype)
  kind_tuple: tuple[str | DTypeLike, ...] = (
    kind if isinstance(kind, tuple) else (kind,)
  )
  options: set[DType] = set()
  for kind in kind_tuple:
    if isinstance(kind, str) and kind in _dtype_kinds:
      options.update(_dtype_kinds[kind])
      continue
    try:
      _dtype = np.dtype(kind)
    except TypeError as e:
      if isinstance(kind, str):
        raise ValueError(
          f"Unrecognized {kind=} expected one of {list(_dtype_kinds.keys())}, "
          "or a compatible input for jnp.dtype()")
      raise TypeError(
        f"Expected kind to be a dtype, string, or tuple; got {kind=}"
      ) from e
    options.add(_dtype)
  return the_dtype in options


def _jax_type(dtype: DType, weak_type: bool) -> JAXType:
  """Return the jax type for a dtype and weak type."""
  if weak_type:
    if dtype == bool:
      return dtype
    if dtype in _custom_float_dtypes:
      return float
    return type(dtype.type(0).item())
  return dtype

View on GitHub (pinned to 1e1c6a8fc0)

Solutions

1. Use exact Array API kind strings: 'bool', 'signed integer', 'unsigned integer', 'real floating', 'complex floating', 'numeric', 'integral'
2. For simple checks prefer jnp.issubdtype(dtype, jnp.integer) / np.dtype comparisons
3. Pass dtype-like objects (np.int32, 'int32') instead of ambiguous short names

Example fix

# before
isdtype(dtype, kind='int')

# after
isdtype(dtype, kind='signed integer')  # or isdtype(dtype, np.int64)

valid_kinds = {'bool','signed integer','unsigned integer','integral','real floating','complex floating','numeric'}
if isinstance(kind, str) and kind not in valid_kinds:
    kind = {'int': 'signed integer', 'uint': 'unsigned integer',
            'float': 'real floating', 'complex': 'complex floating'}.get(kind, kind)

def is_valid_kind(k) -> bool:
    valid = {'bool','signed integer','unsigned integer','integral','real floating','complex floating','numeric'}
    return (isinstance(k, tuple) and all(is_valid_kind(x) for x in k)) or k in valid or _is_dtype_like(k)

try:
    isdtype(dtype, kind)
except ValueError:
    isdtype(dtype, 'signed integer')  # mapped kind

When it happens

Trigger: isdtype(dtype, kind='int') — 'int' is not a valid kind (valid: 'signed integer', 'unsigned integer', 'real floating', etc.); or passing an arbitrary object as kind.

Common situations: Users abbreviating kind names ('float' instead of 'real floating', 'int' instead of 'signed integer') when porting numpy-style checks to the Array API.

Related errors

• dtype cannot be None.
• unexpected input: {dtype=}
• Dtype {dtype} is not a valid JAX array type. Only arrays of
• Value '{x}' with dtype {dt} is not a valid JAX array type. O
• JAX only supports number, bool, and string dtypes, got dtype

AI-assisted analysis of jax-ml/jax@1e1c6a8fc0 (2026-08-27). Data as JSON: /api/errors/61eddf5b3f6c3e63. Report an issue: GitHub.