ba96e585d2ac8e1f940080d5b669976bdec8723b,agents/agent.py,Agent,update_log,#Agent#,153

Before Change

        if phase == RunPhase.TRAIN:
            logger.create_signal_value("Training Reward", self.total_reward_in_current_episode)
        elif phase == RunPhase.TEST:
            logger.create_signal_value("Evaluation Reward", self.total_reward_in_current_episode)
        logger.update_wall_clock_time(self.current_episode)

        for signal in self.signals:
            logger.create_signal_value("{}/Mean".format(signal.name), signal.get_mean())

After Change

        logger.create_signal_value("Epsilon", self.exploration_policy.get_control_param())
        logger.create_signal_value("Training Reward", self.total_reward_in_current_episode
                                   if phase == RunPhase.TRAIN else np.nan)
        logger.create_signal_value("Evaluation Reward", self.total_reward_in_current_episode
                                   if phase == RunPhase.TEST else np.nan)
        logger.create_signal_value("Update Target Network", 0, overwrite=False)
        logger.update_wall_clock_time(self.current_episode)

        for signal in self.signals:

Instances