# How does training & evaluation work in CrewAI?

**URL:** https://community.crewai.com/t/how-does-training-evaluation-work-in-crewai/4160
**Category:** General
**Created:** [February 25, 2025, 3:55am UTC](https://community.crewai.com/t/how-does-training-evaluation-work-in-crewai/4160 "2025-02-25T03:55:54Z")
**Posts on this page:** 1
**Page:** 1

<div class="post-metadata">

### Author: ![pratikchhapolika](https://sea1.discourse-cdn.com/flex025/user_avatar/community.crewai.com/pratikchhapolika/32/2434_2.png) [@pratikchhapolika](https://community.crewai.com/u/pratikchhapolika)
#### Post date: [February 25, 2025, 3:55am UTC](https://community.crewai.com/t/how-does-training-evaluation-work-in-crewai/4160/1 "2025-02-25T03:55:54Z")

</div>

I went through this post: [How does training & evaluation actually work in CrewAI?](https://community.crewai.com/t/how-does-training-evaluation-actually-work-in-crewai/1341)

But I don’t think its that easy.

If just the prompts are injected in next run then why its called a training?

I see 2 `.pkl` files are saved `training_data.pkl` and `usecase.pkl`

Does it update some weights? I would like to understand step by step of how training happens. I assume we are not updating the weights of underlying LLM models as those might be big gpt models.

@matt