# Not Recording MultiAgentEnv on Evaluation

**URL:** <https://discuss.ray.io/t/not-recording-multiagentenv-on-evaluation/12912>\
**Category:** RLlib\
**Created:** [November 22, 2023, 5:00pm UTC](https://discuss.ray.io/t/not-recording-multiagentenv-on-evaluation/12912 "2023-11-22T17:00:35Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![Brendan\_A](https://sea2.discourse-cdn.com/flex020/user_avatar/discuss.ray.io/brendan_a/32/110_2.png) [@Brendan\_A](https://discuss.ray.io/u/Brendan_A)\
**Post date:** [November 22, 2023, 5:00pm UTC](https://discuss.ray.io/t/not-recording-multiagentenv-on-evaluation/12912/1 "2023-11-22T17:00:35Z")

</div>

I have a custom `MultiAgentEnv` that I am working with right now that I can get to render standalone with my own evaluation loop. But I can not get it to record the videos. And when I try to use the `AlgorithmConfig().evaluation()` method I can’t get the env to render or record. I’ve poked around a number of the related topics on this but can not seem to get anything to quite work as expected.

The things I have made sure of so far:

- the MultiAgentEnv.metadata attribute includes `"render.modes":["rgb_array","human"]`
- the MultiAgentEnv contains it’s own `render()` method
- when I create the env\_config for the manual evaluation loop, I’ve included both `"record_env":True` and `"monitor":True`
- when I try to record using the `AlgorithmConfig().evaluation()`, I include the argument `evaluation_config={"record_env":True, "render_env":True}`

I’m kinda at a loss as to what else I should try to get the env to record. Any suggestions are appreciated.

- Medium: It contributes to significant difficulty to complete my task, but I can work around it.