Remove assign ops, compute_loss

      return int(self._global_step)

    return self._global_step.eval(session=self.session)

  def compute_loss(self, dataset, transformers):

    if self.mode == "regression":

      y = undo_transforms(dataset.y, transformers=transformers)

    else:

      y = to_one_hot(dataset.y.flatten(), self.n_classes).reshape(

          -1, len(self.n_tasks), self.n_classes)

    loss_fn = model._loss_fn

    y_pred = self.predict(dataset, transformers=transformers)

    loss_tensor = loss_fn([y_pred], [y], [dataset.w])

    if tf.executing_eagerly():

      return loss_tensor.numpy()

    else:

      return self.session.run(loss_tensor)

  def _create_assignment_map(self, source_model, include_top=True, **kwargs):

    """

    Creates a default assignment map between variables of source and current model.