jax-ml/jax · error · ValueError

Unknown FFT type '{s}'

Error message

Unknown FFT type '{s}'

What it means

jax.lax.fft accepts the FFT type either as a FftType enum or as one of the strings 'fft','ifft','rfft','irfft' (case-insensitive). Any other string falls through _str_to_fft_type and raises this ValueError echoing the unknown name.

Source

Thrown at jax/_src/lax/fft.py:65

  RFFT = 2
  "Forward real-to-complex FFT."

  IRFFT = 3
  "Inverse real-to-complex FFT."


def _str_to_fft_type(s: str) -> FftType:
  if s in ("fft", "FFT"):
    return FftType.FFT
  elif s in ("ifft", "IFFT"):
    return FftType.IFFT
  elif s in ("rfft", "RFFT"):
    return FftType.RFFT
  elif s in ("irfft", "IRFFT"):
    return FftType.IRFFT
  else:
    raise ValueError(f"Unknown FFT type '{s}'")

@jit(static_argnums=(1, 2))
def fft(x, fft_type: FftType | str, fft_lengths: Sequence[int]):
  if isinstance(fft_type, str):
    typ = _str_to_fft_type(fft_type)
  elif isinstance(fft_type, FftType):
    typ = fft_type
  else:
    raise TypeError(f"Unknown FFT type value '{fft_type}'")

  if typ == FftType.RFFT:
    if np.iscomplexobj(x):
      raise ValueError("only real valued inputs supported for rfft")
    x = lax.convert_element_type(x, dtypes.to_inexact_dtype(dtypes.dtype(x)))
  else:
    x = lax.convert_element_type(x, dtypes.to_complex_dtype(dtypes.dtype(x)))
  if len(fft_lengths) == 0:
    # XLA FFT doesn't support 0-rank.

View on GitHub (pinned to 1e1c6a8fc0)

Solutions

1. Use one of 'fft', 'ifft', 'rfft', 'irfft' or the lax.FftType enum members
2. For multi-axis FFTs, pass all axes in fft_lengths with type 'fft' rather than a 'fft2' name
3. Prefer jax.numpy.fft wrappers (jnp.fft.fft2 etc.) for familiar names

Example fix

# before
y = lax.fft(x, 'fft2', (64, 64))
# after
y = lax.fft(x, 'fft', (64, 64))  # or jnp.fft.fft2(x)

assert fft_type in ('fft','ifft','rfft','irfft') or isinstance(fft_type, lax.FftType)

def valid_fft_type(t) -> bool:
    return (isinstance(t, lax.FftType) or
            (isinstance(t, str) and t.lower() in ('fft','ifft','rfft','irfft')))

When it happens

Trigger: Calling jax.lax.fft(x, 'dft', (n,)) or 'fft2'/'complex_fft' — names that exist in other libraries but not here.

Common situations: Porting from numpy/scipy or TensorFlow naming (fft2, fftn, hfft); typos like 'fftn' expecting multi-axis behavior.

Understand the failure class

Background: "Must be a positive integer", "Invalid value", "Unsupported": the invalid-argument-value error family, when a library rejects the value you pass — this error's family across 35 libraries.

Related errors

• only real valued inputs supported for rfft
• {name} wrapped function must be passed at least one argument
• primal and tangent arguments to jax.jvp must be tuples or li
• check_error takes an Error as argument, got type {type(error
• Can't use ``defjvps`` with ``nondiff_argnums``.

AI-assisted analysis of jax-ml/jax@1e1c6a8fc0 (2026-08-27). Data as JSON: /api/errors/76639c5959a4bb4a. Report an issue: GitHub.