# Finetuning MBMPO policy

**URL:** <https://discuss.ray.io/t/finetuning-mbmpo-policy/7530>\
**Category:** RLlib\
**Created:** [September 13, 2022, 1:07am UTC](https://discuss.ray.io/t/finetuning-mbmpo-policy/7530 "2022-09-13T01:07:26Z")\
**Posts on this page:** 1\
**Showing post:** 3

<div class="post-metadata">

**Author:** ![Nehal\_Soni](https://sea2.discourse-cdn.com/flex020/user_avatar/discuss.ray.io/nehal_soni/32/1706_2.png) [@Nehal\_Soni](https://discuss.ray.io/u/Nehal_Soni)\
**Post date:** [September 13, 2022, 7:31pm UTC](https://discuss.ray.io/t/finetuning-mbmpo-policy/7530/3 "2022-09-13T19:31:04Z")

</div>

Thank you @arturn for your quick response, it’s helpful.

I understand that MAML policy needs to be fine-tuned and it is possible directly using PPO algorithm of RLlib (This thread mentions it and it has been tested also: [MAML finetune adaptation step for inference](https://discuss.ray.io/t/maml-finetune-adaptation-step-for-inference/4651)).

Is there any better approach in RLlib to fine-tune MBMPO policy?

Regards

---

_[View the full topic](https://discuss.ray.io/t/finetuning-mbmpo-policy/7530)._