jax-ml/jax · error · ValueError

Unrecognized method {method}. The two valid methods are eith

Error message

Unrecognized method {method}. The two valid methods are either \"schur\" or \"eigen\".

What it means

jax.scipy.linalg.solve_sylvester supports only method='schur' or method='eigen'; anything else reaches the else-branch raising this ValueError. The method selects the Bartels-Stewart Schur decomposition or eigendecomposition path.

Source

Thrown at jax/_src/scipy/linalg.py:3291

@jit(static_argnames=["method"])
def _solve_sylvester_2d(A: Array, B: Array, C: Array, *, method: str, tol: float) -> Array:
  m, n = C.shape[-2:]
  if method == "schur":
    R, U = schur(A, output='complex')
    S, V = schur(B.conj().T, output='complex')
    F = U.conj().T @ C.astype(R.dtype) @ V
    Y = _solve_sylvester_triangular_scan(R, S.conj().T, F)
    X = U @ Y @ V.conj().T
  elif method == "eigen":
    RA, UA = jnp.linalg.eig(A)
    RB, UB = jnp.linalg.eig(B)
    F = solve(UA, C.astype(RA.dtype) @ UB)
    W = RA[:, None] + RB[None, :]
    Y = F / W
    X = UA[:m,:m] @ Y[:m,:n] @ inv(UB)[:n,:n]
  else:
    raise ValueError(f"Unrecognized method {method}. The two valid methods are either \"schur\" or \"eigen\".")
  if not dtypes.issubdtype(C.dtype, np.complexfloating):
    X = X.real
  return lax.cond(
    jnp.any(jnp.abs(jnp.linalg.eigvals(A)[:, None] + jnp.linalg.eigvals(B)[None, :]) < tol),
    lambda: jnp.zeros_like(X) * np.nan,
    lambda: X,
  )


@jit(static_argnames=["method"])
def solve_sylvester(A: ArrayLike, B: ArrayLike, C: ArrayLike, *, method: str = "schur", tol: float = 1e-8) -> Array:
  """
  Solves the Sylvester equation
  .. math::

    AX + XB = C

  Using one of two methods.

View on GitHub (pinned to 1e1c6a8fc0)

Solutions

1. Use 'schur' (default, more stable) or 'eigen'
2. Validate method strings before calling

Example fix

# before
X = linalg.solve_sylvester(A, B, C, method='schur ')  # stray space
# after
X = linalg.solve_sylvester(A, B, C, method='schur')

if method not in ('schur','eigen'): raise ValueError(method)

def is_valid_method(m: str) -> bool: return m in ('schur', 'eigen')

When it happens

Trigger: Calling solve_sylvester(A, B, C, method='qr'), method='', or a misspelled string like 'Schur'.

Common situations: Copy-paste from custom solvers; passing a method intended for a different scipy function.

Related errors

• ind must be a positive integer; got {ind=}
• Expected kind to be on of: {valid_kind}; got {kind}
• Expected kind to be one of: {valid_kind}; got {kind}
• n must be a positive power of 2; got {n}.
• scale must be None, 'sqrtn', or 'n'; got {scale!r}.

AI-assisted analysis of jax-ml/jax@1e1c6a8fc0 (2026-08-27). Data as JSON: /api/errors/29692bb3f5c483f2. Report an issue: GitHub.