86ce5d52134a56806112ff8664e4034338e0e05a,yarll/agents/ppo/ppo.py,PPO,learn,#PPO#,208

Before Change

                    batch_rs = np.array(rs)[batch_indices]
                    fetches = [self.train_op]
                    if (n_updates % 1000) == 0:
                        fetches.append(self.model_summary_op)
                    feed_dict = {
                        self.states: batch_states,
                        self.old_network.states: batch_states,
                        self.actions_taken: batch_actions,

After Change

                                                                                     batch_actions,
                                                                                     batch_advs,
                                                                                     batch_rs)
                        tf.summary.scalar("model/loss", train_loss, step=n_updates)
                        tf.summary.scalar("model/actor_loss", train_actor_loss, step=n_updates)
                        tf.summary.scalar("model/critic_loss", train_critic_loss, step=n_updates)
                        n_updates += 1

Instances