def _apply_sparse(self, grad, var): acc = self.get_slot(var, "accumulator") return training_ops.sparse_apply_proximal_adagrad( var, acc, self._learning_rate_tensor, self._l1_regularization_strength_tensor, self._l2_regularization_strength_tensor, grad.values, grad.indices, use_locking=self._use_locking)
def _apply_sparse(self, grad, var): acc = self.get_slot(var, "accumulator") return training_ops.sparse_apply_proximal_adagrad( var, acc, self._learning_rate_tensor, self._l1_regularization_strength_tensor, self._l2_regularization_strength_tensor, grad.values, grad.indices, use_locking=self._use_locking)