75dff38e70ec291fbd60382174277fcadddfd285,reagent/ope/estimators/contextual_bandits_estimators.py,DMEstimator,evaluate,#DMEstimator#Any#,82

Before Change

            self._append_estimate(
                log_reward.average, tgt_reward.average, gt_reward.average
            )
        return self.results


class IPSEstimator(Estimator):
    

After Change

            _, tgt_reward = self._calc_dm_reward(input.action_space, sample)
            tgt_avg.add(tgt_reward)
            gt_avg.add(sample.ground_truth_reward)
        return EstimatorResult(
            log_avg.average, tgt_avg.average, gt_avg.average, tgt_avg.count
        )

    def __repr__(self):
        return f"DMEstimator(trainer({self._trainer.name},device({self._device}))"

Instances