# Submission format of Descriptor Track

**URL:** <https://community.drivendata.org/t/submission-format-of-descriptor-track/6248>\
**Category:** Image Similarity Challenge\
**Created:** [July 4, 2021, 1:07pm UTC](https://community.drivendata.org/t/submission-format-of-descriptor-track/6248 "2021-07-04T13:07:01Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![songgogogo](https://avatars.discourse-cdn.com/v4/letter/s/87869e/32.png) [@songgogogo](https://community.drivendata.org/u/songgogogo)\
**Post date:** [July 4, 2021, 1:07pm UTC](https://community.drivendata.org/t/submission-format-of-descriptor-track/6248/1 "2021-07-04T13:07:01Z")

</div>

On the Problem description page, I submit a single HDF5 file containing four datasets(query, reference, query\_ids, reference\_ids) . But it prompts me to submit .csv file. I am confused about this！  
And there is an submission error when I submit a .h5 file.

---

<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:** [July 6, 2021, 6:08pm UTC](https://community.drivendata.org/t/submission-format-of-descriptor-track/6248/2 "2021-07-06T18:08:53Z")

</div>

Hi songgogogo! Thanks for the question, and apologies for the confusion. The submission page has been updated to ask for an HDF5 file instead of a csv file.

It should be accepting valid HDF5 files despite the incorrect message. Keep us posted if you are still getting submission errors after submitting a valid file, as detailed on [the Problem Description page](https://www.drivendata.org/competitions/80/competition-image-similarity-2-dev/page/380/#submissionformat).