All episodes SAME REWARD!

[skhonsha]

Dears, Thank you for framework, Please see the output of hyperparameter training on SB3 algorithm, why the reward in all episodes doesnt change, whats the problem?(I copied only three outputs)The reward output is the same in every study for any hyperparameter configuration

---

| time/ | | | episodes | 172 | | fps | 92 | | time_elapsed | 1014 | | total_timesteps | 93568 | | train/ | | | actor_loss | 5.13 | | critic_loss | 7.43 | | ent_coef | 0.00201 | | learning_rate | 0.0148 | | n_updates | 93447 | | reward | -3.8389 |

---

| time/ | | | episodes | 176 | | fps | 92 | | time_elapsed | 1039 | | total_timesteps | 95744 | | train/ | | | actor_loss | 4.84 | | critic_loss | 39.8 | | ent_coef | 0.00201 | | learning_rate | 0.0148 | | n_updates | 95623 | | reward | -3.8389 |

---

| time/ | | | episodes | 180 | | fps | 92 | | time_elapsed | 1062 | | total_timesteps | 97920 | | train/ | | | actor_loss | 4.44 | | critic_loss | 3.66 | | ent_coef | 0.00201 | | learning_rate | 0.0148 | | n_updates | 97799 | | reward | -3.8389 |

[YangletLiu]

@skhonsha Please give more details, since we did not encounter it before. Which code file are you using？

[KoperSloper]

Have you been able to find a solution?