d670e94d3f8e938858e938b1d3efeac5389a26d3,slm_lab/experiment/control.py,Session,run_episode,#Session#,45

Before Change

                break
        logger.info(f"epi {self.aeb_space.clock["e"]}, total_rewards {total_rewards}")
        // TODO compose episode data properly with monitor
        episode_data = pd.DataFrame(
            episode_data_list, columns=["rewards", "total_rewards", "loss", "explore_var"])
        // episode_data = {}
        return episode_data

    def run(self):

After Change

            // if bool(done_space):
            if done_space.get(a=0)[0]:
                // TODO refactor: set all to terminate on master termination. Also use the env with longest timestep to prevent being terminated by fast-running env
                for a, _eb in enumerate(self.aeb_space.a_eb_proj):
                    done_proj_a = done_space.get(a=a)
                    // TODO still need to standardize all data proj to aeb
                    for a_idx, _done in enumerate(done_proj_a):
                        done_proj_a[a_idx] = True
                break
        // TODO monitor record all data spaces, including body with body.clock. cuz all data spaces have history
        // split per body, use done as delim (maybe done need body clock now), split, sum each chunk
        mean_loss = np.nanmean(loss_list)

Instances