ut3_zeros
=========

.. py:function:: t3toolbox.backend.ut3_constructors.ut3_zeros(shape, tucker_ranks=None, tt_ranks=None, stack_shape = (), use_jax = False)

   .. code-block:: python

      def ut3_zeros(
              shape:        typ.Sequence[int],               # (N0,...,N(d-1))
              tucker_ranks=None,                             # int | len-d seq | (d,)+stack array;   None -> all 1
              tt_ranks=None,                                 # int | len-(d+1) seq | (d+1,)+stack array; None -> all 1
              stack_shape:  typ.Tuple[int, ...] = (),
              use_jax:      bool = False,                    # constructor: chooses the SUPERCORE type
      ) -> UT3Data:


   Uniform Tucker tensor train of zeros. ``tucker_ranks``/``tt_ranks`` may vary per stack element.