# Ground Truth for sample submission

**URL:** <https://community.drivendata.org/t/ground-truth-for-sample-submission/7765>\
**Category:** Where's Whale-do?\
**Created:** [May 31, 2022, 12:20pm UTC](https://community.drivendata.org/t/ground-truth-for-sample-submission/7765 "2022-05-31T12:20:17Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![Pantazis](https://avatars.discourse-cdn.com/v4/letter/p/43a26b/32.png) [@Pantazis](https://community.drivendata.org/u/Pantazis)\
**Post date:** [May 31, 2022, 12:20pm UTC](https://community.drivendata.org/t/ground-truth-for-sample-submission/7765/1 "2022-05-31T12:20:17Z")

</div>

Hi everyone,  
If we want to test our model locally with a sample submission we have created, how must our ground truth csv be created? I know the photos of the same whale must be included for sure, but what about the rest of the twenty (if the same whale id appears in the database fewer times than 20) ?

Thanks in advance!

---

<div class="post-metadata">

**Author:** ![mike-dd](https://avatars.discourse-cdn.com/v4/letter/m/f0a364/32.png) [@mike-dd](https://community.drivendata.org/u/mike-dd)\
**Post date:** [June 1, 2022, 3:20pm UTC](https://community.drivendata.org/t/ground-truth-for-sample-submission/7765/2 "2022-06-01T15:20:30Z")

</div>

Hi @Pantazis,

Please check out [this section](https://www.drivendata.org/competitions/96/beluga-whales/page/482/#scoring_ground_truth) of the competition website, which describes a script to use for local scoring and the format for the ground truth CSV file.

I hope this helps.