a4cac743f6094a5f9168a2b31c9481ace2d9957f,tensorforce/rl_agents/pg_agent.py,PGAgent,add_observation,#PGAgent#Any#Any#Any#Any#,69

Before Change

        :param terminal:
        :return:
        
        state = self.filter(state)
        self.batch_steps += 1
        self.current_episode["states"].append(state)
        self.current_episode["actions"].append(self.last_action)
        self.current_episode["rewards"].append(reward)

After Change

            // Transform into np arrays, append episode to batch, start new episode dict
            path = self.get_path()
            self.current_batch.append(path)
            print("Episode finished, length=" + str(len(path["states"])))
            self.current_episode = defaultdict(list)

        if self.batch_steps == self.batch_size:

Instances