jax-ml/jax · error · ValueError

b matrix must have rank >= 1, got shape {b.shape}

Error message

b matrix must have rank >= 1, got shape {b.shape}

What it means

jax/_src/lax/linalg.py:1833 in _lu_solve (public lu_solve). The right-hand side b must be at least rank 1 (a vector). A 0-d scalar b has no axis to solve along, so the ValueError fires immediately before broadcasting logic runs.

Source

Thrown at jax/_src/lax/linalg.py:1833

    conj = trans == 2
    x = triangular_solve(lu, x, left_side=True, lower=False, transpose_a=True,
                         conjugate_a=conj)
    x = triangular_solve(lu, x, left_side=True, lower=True, unit_diagonal=True,
                         transpose_a=True, conjugate_a=conj)
    _, ind = lax.sort_key_val(permutation, lax.iota('int32', permutation.shape[0]))
    x = x[ind, :]
  else:
    raise ValueError(f"'trans' value must be 0, 1, or 2, got {trans}")
  return lax.reshape(x, b.shape)


@api.jit(static_argnums=(3,))
def _lu_solve(lu: Array, permutation: Array, b: Array, trans: int) -> Array:
  if len(lu.shape) < 2 or lu.shape[-1] != lu.shape[-2]:
    raise ValueError("last two dimensions of LU decomposition must be equal, "
                     "got shape {}".format(lu.shape))
  if len(b.shape) < 1:
    raise ValueError("b matrix must have rank >= 1, got shape {}"
                     .format(b.shape))
  # Broadcasting follows NumPy's convention for linalg.solve: the RHS is
  # treated as a (batched) vector if the number of dimensions differ by 1.
  # Otherwise, broadcasting rules apply.
  rhs_vector = lu.ndim == b.ndim + 1
  if rhs_vector:
    if b.shape[-1] != lu.shape[-1]:
      raise ValueError("When LU decomposition matrix and b have the same "
                       "number of dimensions, last axis of LU decomposition "
                       "matrix (shape {}) and b array (shape {}) must match"
                       .format(lu.shape, b.shape))
    b = b[..., np.newaxis]
  else:
    if b.shape[-2] != lu.shape[-1]:
      raise ValueError("When LU decomposition matrix and b different "
                       "numbers of dimensions, last axis of LU decomposition "
                       "matrix (shape {}) and second to last axis of b array "
                       "(shape {}) must match"

View on GitHub (pinned to 1e1c6a8fc0)

Solutions

1. Keep b at least 1-d: b = jnp.atleast_1d(b) or b = b[None]
2. Replace accidental reductions (e.g. use keepdims=True on sums feeding b)
3. Add an assert b.ndim >= 1 during development

Example fix

// before
x = jax.lax.linalg.lu_solve(lu, piv, b)  # b is 0-d scalar
// after
x = jax.lax.linalg.lu_solve(lu, piv, jnp.atleast_1d(b))

b = jnp.atleast_1d(b)

def valid_rhs(b: jax.Array) -> bool:
    return b.ndim >= 1

When it happens

Trigger: Calling jax.lax.linalg.lu_solve(lu, permutation, b) with b = jnp.scalar or a Python float, e.g. accidentally reducing b with .sum() or indexing b[i] to a scalar before passing.

Common situations: Loop-refactor bugs where b[i] should be b[i:i+1]; aggressive squeezing (jnp.squeeze) collapsing a (1,) RHS to 0-d; mixing scalar coefficients with matrix solves in solvers for ODE roots.

Related errors

• last two dimensions of LU decomposition must be equal, got s
• When LU decomposition matrix and b have the same number of d
• When LU decomposition matrix and b different numbers of dime
• {name} was requested to map its argument along axis {axis},
• The process id of the current process must be defined.

AI-assisted analysis of jax-ml/jax@1e1c6a8fc0 (2026-08-27). Data as JSON: /api/errors/9f037d2a14fd3af7. Report an issue: GitHub.