# Reinforcement learning model doesn't improve

**URL:** <https://discuss.ai.google.dev/t/reinforcement-learning-model-doesnt-improve/28618>\
**Category:** General Discussion\
**Tags:** reinforcement-learni\
**Created:** [April 17, 2023, 2:04am UTC](https://discuss.ai.google.dev/t/reinforcement-learning-model-doesnt-improve/28618 "2023-04-17T02:04:10Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![Andrija\_Gajic](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/andrija_gajic/32/15244_2.png) [@Andrija\_Gajic](https://discuss.ai.google.dev/u/Andrija_Gajic)\
**Post date:** [April 17, 2023, 2:04am UTC](https://discuss.ai.google.dev/t/reinforcement-learning-model-doesnt-improve/28618/1 "2023-04-17T02:04:10Z")

</div>

I am training the model that plays dino chrome game. I’ve used DQN algorithm and puppeteer for automation. I ran the code for 5000 episodes but it doesn’t seem to improve. Can somebody take a look a the code, here is the [github link](https://github.com/gajojr/tfjs-dino-player)