# Chapter 5: Hyper parameters on Turi create image\_classifier

**URL:** <https://forums.kodeco.com/t/chapter-5-hyper-parameters-on-turi-create-image-classifier/57003>\
**Category:** Machine Learning by Tutorials\
**Created:** [December 5, 2018, 11:42am UTC](https://forums.kodeco.com/t/chapter-5-hyper-parameters-on-turi-create-image-classifier/57003 "2018-12-05T11:42:29Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![scy](https://assets.chunter.kodeco.com/user_avatar/forums.kodeco.com/scy/32/105896_2.png) [@scy](https://forums.kodeco.com/u/scy)\
**Post date:** [December 5, 2018, 11:42am UTC](https://forums.kodeco.com/t/chapter-5-hyper-parameters-on-turi-create-image-classifier/57003/1 "2018-12-05T11:42:29Z")

</div>

on chapter 5 on the book , it is written that you can decrease the amount of overfitting by changing some hyper parameters.

Is it possible to get a concrete example how to do that ?

Thanks

---

<div class="post-metadata">

**Author:** ![hollance](https://assets.chunter.kodeco.com/user_avatar/forums.kodeco.com/hollance/32/109009_2.png) [@hollance](https://forums.kodeco.com/u/hollance)\
**Post date:** [December 5, 2018, 1:05pm UTC](https://forums.kodeco.com/t/chapter-5-hyper-parameters-on-turi-create-image-classifier/57003/2 "2018-12-05T13:05:20Z")

</div>

In the section “Training the classifier with regularization” of that chapter, it shows how to change the L1 / L2 penalty to reduce overfitting. That is an example of changing the hyperparameters.