# Run custom environment with episodes-based replay buffer slots

**URL:** <https://discuss.ray.io/t/run-custom-environment-with-episodes-based-replay-buffer-slots/7211>\
**Category:** RLlib\
**Created:** [August 16, 2022, 9:23am UTC](https://discuss.ray.io/t/run-custom-environment-with-episodes-based-replay-buffer-slots/7211 "2022-08-16T09:23:03Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![Manna](https://sea2.discourse-cdn.com/flex020/user_avatar/discuss.ray.io/manna/32/3014_2.png) [@Manna](https://discuss.ray.io/u/Manna)\
**Post date:** [August 16, 2022, 9:23am UTC](https://discuss.ray.io/t/run-custom-environment-with-episodes-based-replay-buffer-slots/7211/1 "2022-08-16T09:23:03Z")

</div>

**How severe does this issue affect your experience of using Ray?**  
High: It blocks me to complete my task.

Hello,

I am using a custom environment and I would like to investigate the effects of using sequences and episodes in the replay buffer on my training convergence.  
I have changed the ‘replay\_buffer\_config’ to reflect this but I am getting errors as soon as I am changing the ‘storage\_unit’ field.  
For example, when I set storage\_unit to episodes, the error says that I am not defining the ‘T’ field used by the replay buffer when using this specific storage unit. Similar error when I choose ‘sequences’ as a storage unit, I am missing ‘SEQ\_LENS’.  
Do I need to add any additional output to my environment, or what other settings should I change in the configuration to get it to work?  
Thanks!