jax-ml/jax · error · ValueError

x.shape={x.shape} != mask.shape={mask.shape}

Error message

x.shape={x.shape} != mask.shape={mask.shape}

What it means

scan_count requires mask.shape == x.shape exactly; broadcasting is not applied.

Source

Thrown at jax/_src/pallas/mosaic/sc_primitives.py:531

  The barrier must be used with
  :class:`jax.experimental.pallas.tpu_sc.VectorSubcoreMesh`.
  """
  barrier_p.bind()


scan_count_p = jax_core.Primitive("scan_count")
scan_count_p.multiple_results = True


@scan_count_p.def_abstract_eval
def _scan_count_abstract_eval(x, mask):
  if x.dtype not in (jnp.uint32, jnp.int32, jnp.float32):
    raise NotImplementedError(
        f"x.dtype={x.dtype} must be uint32, int32 or float32")
  if not jnp.issubdtype(mask.dtype, jnp.bool):
    raise TypeError(f"mask.dtype={mask.dtype} is not a boolean dtype")
  if x.shape != mask.shape:
    raise ValueError(f"x.shape={x.shape} != mask.shape={mask.shape}")
  return jax_core.ShapedArray(x.shape, jnp.int32), mask


@sc_lowering.register_lowering_rule(scan_count_p)
def _scan_count_lowering_rule(ctx: sc_lowering.LoweringRuleContext, x, mask):
  del ctx  # Unused.
  # Reverse, because the MLIR op returns the mask first.
  return tpu.scan_count(mask, x)[::-1]


def scan_count(
    x: jax.Array, mask: jax.Array | None = None
) -> tuple[jax.Array, jax.Array]:
  """Computes the running duplicate occurrence count of the array.

  Args:
    x: An array of integers or floats.
    mask: An optional array of booleans, which specifies which elements ``x``

View on GitHub (pinned to 1e1c6a8fc0)

Solutions

1. Broadcast the mask to x.shape explicitly: jnp.broadcast_to(mask, x.shape)
2. Recompute the mask at matching shape inside the kernel

Example fix

// before
scan_count(x, mask)  # mask (N,), x (N,M)

// after
scan_count(x, jnp.broadcast_to(mask, x.shape))

if mask.shape != x.shape:
    mask = jnp.broadcast_to(mask, x.shape)

def shapes_match(x, mask) -> bool:
    return tuple(x.shape) == tuple(mask.shape)

When it happens

Trigger: Passing mask of shape (N,) with x of shape (N, M), or a scalar mask with any x.

Common situations: Assuming mask broadcasts over the vector dimension; reusing a batch-level mask for token-level data.

Related errors

• {x.shape=} does not match expected shape {expected_shape}
• x.dtype={x.dtype} must be uint32, int32 or float32
• mask.dtype={mask.dtype} is not a boolean dtype
• Block shape for {origin} (= {block_shape}) must have the sam
• Stacking only supported when the block size along the stack

AI-assisted analysis of jax-ml/jax@1e1c6a8fc0 (2026-08-27). Data as JSON: /api/errors/5359c6a77f1086aa. Report an issue: GitHub.