dc1835982db065302b92ba8a8fdf4139af910a66,ml/rl/test/gridworld/gridworld_evaluator.py,GridworldDDPGEvaluator,evaluate_critic,#GridworldDDPGEvaluator#Any#,387

Before Change

            actions=self.logged_actions,
        )
        error_sum = 0.0
        for x in range(len(self.logged_states)):
            ground_truth = self.logged_values[x][0]
            target_value = critic_prediction[x]
            error_sum += abs(ground_truth - target_value)
        logger.info(
            "EVAL ERROR: {0:.3f}".format(error_sum / float(len(self.logged_states)))
        )
        self.mc_loss.append(error_sum / float(len(self.logged_states)))

After Change

                )

    def evaluate_critic(self, critic):
        return super(GridworldDDPGEvaluator, self).evaluate(critic)

Instances