// [1]
  auto reduced = Cast(cpu_scope, reduction_indices_pos, DataType::DT_INT32);

  // [0, 1, 2]
  auto idx = Range(cpu_scope, zero, rank, one);

  // [0, 2]
  auto other = SetDiff1D(cpu_scope, idx, reduced).out;

  // [1, 0, 2]
  auto perm =
      Concat(cpu_scope, std::initializer_list<Input>{reduced, other}, 0);

  // 3 => [3]

`tf.sparse_reduce_sum`: `reduction_axes` becomes `axis` *
`tf.sparse_reduce_sum_sparse`: `reduction_axes` becomes `axis` *
`tf.sparse_split`: `split_dim` becomes `axis` * `tf.listdiff` has been renamed
to `tf.setdiff1d` to match NumPy naming. * `tf.inv` has been renamed to be
`tf.reciprocal` (component-wise reciprocal) to avoid confusion with `np.inv`
which is matrix inversion * tf.round now uses banker's rounding (round to even)