utv_entries_transpose
=====================

.. py:function:: t3toolbox.backend.utv_sampling.utv_entries_transpose(c, index, frame_data, sum_over_probes=False)

   .. code-block:: python

      def utv_entries_transpose(
              c,                # residual, shape=W+K+C
              index,            # int array, shape=(d,)+W -- the indices c weights
              frame_data,       # UT3Frame .data (orthogonal frame), stack=C
              sum_over_probes=False,
      ):  # -> UT3Variations .data: (dU_tilde, dG_tilde, shape, masks); stack K_new + C


   Apply the transpose ``𝒥ᵀ`` of the all-modes entries (scatter ``c`` at ``index``). Like
   :py:func:`utv_apply_transpose` with the apply vectors replaced by the one-hot ``e_index`` (built
   packed inside ``entries._onehot_vectors``), so ``index`` needs no packing.