# Multi\_scale parameter Yolov8

**URL:** https://community.ultralytics.com/t/multi-scale-parameter-yolov8/1776
**Category:** Discussion
**Tags:** question, troubleshooting, yolo, support
**Created:** [January 27, 2026, 9:02pm UTC](https://community.ultralytics.com/t/multi-scale-parameter-yolov8/1776 "2026-01-27T21:02:14Z")
**Posts on this page:** 2
**Page:** 1

<div class="post-metadata">

### Author: ![Riccardo\_Carraro](https://sea1.discourse-cdn.com/flex001/user_avatar/community.ultralytics.com/riccardo_carraro/32/1150_2.png) [@Riccardo\_Carraro](https://community.ultralytics.com/u/Riccardo_Carraro)
#### Post date: [January 27, 2026, 9:02pm UTC](https://community.ultralytics.com/t/multi-scale-parameter-yolov8/1776/1 "2026-01-27T21:02:15Z")

</div>

I have a question: how does the multi\_scale parameter work? I’m currently working with Yolov8n and i need to make the model “scale invariant” for a particular task of object detection (similar to QR detection where sometimes QR are really big, and sometimes really small)

Reading the documentation, I’ve seen that there is a “multi\_scale” parameter which is a float, BUT, if I put a float value, it says “Boolean required”.

So, if i put “multi\_scale = True” how does the mutliscale work? what it really does?

---

<div class="post-metadata">

### Author: ![Toxite](https://sea1.discourse-cdn.com/flex001/user_avatar/community.ultralytics.com/toxite/32/123_2.png) [@Toxite](https://community.ultralytics.com/u/Toxite)
#### Post date: [January 27, 2026, 11:35pm UTC](https://community.ultralytics.com/t/multi-scale-parameter-yolov8/1776/2 "2026-01-27T23:35:25Z")

</div>

`multi_scale` changes the image size. It is different from `scale` that zooms in and out the image. You should use `scale.`It’s enabled by default.

> **[Augmentation Settings and Hyperparameters - Model Training with Ultralytics YOLO](https://docs.ultralytics.com/modes/train/#augmentation-settings-and-hyperparameters)**
>
> Augmentation techniques are essential for improving the robustness and performance of YOLO models by introducing variability into the training data, helping the model generalize better to unseen data. The following table outlines the purpose and...

> [@Riccardo\_Carraro](#):
>
> Reading the documentation, I’ve seen that there is a “multi\_scale” parameter which is a float, BUT, if I put a float value, it says “Boolean required”.

It used to be boolean value. Now it’s float. You’re likely not using the latest Ultralytics package, hence the error.
