f75a44cac3ae7bfc8810bad5127854a131d48a9c,reagent/ope/estimators/contextual_bandits_estimators.py,DoublyRobustEstimator,evaluate,#DoublyRobustEstimator#Any#,257

Before Change

    def evaluate(
        self, input: BanditsEstimatorInput, **kwargs
    ) -> Optional[EstimatorResult]:
        logger = Estimator.logger()
        self._train_model(input.samples, 0.8, logger)
        log_avg = RunningAverage()
        tgt_avg = RunningAverage()
        gt_avg = RunningAverage()

After Change

        log_avg = RunningAverage()
        logged_vals = []
        tgt_avg = RunningAverage()
        tgt_vals = []
        gt_avg = RunningAverage()
        for sample in input.samples:
            log_avg.add(sample.log_reward)
            logged_vals.append(sample.log_reward)

Instances