# Using Pre-trained models

**URL:** <https://community.drivendata.org/t/using-pre-trained-models/6178>\
**Category:** Overhead Geopose Challenge\
**Created:** [June 10, 2021, 10:02am UTC](https://community.drivendata.org/t/using-pre-trained-models/6178 "2021-06-10T10:02:22Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![tetelias](https://avatars.discourse-cdn.com/v4/letter/t/8e7dd6/32.png) [@tetelias](https://community.drivendata.org/u/tetelias)\
**Post date:** [June 10, 2021, 10:02am UTC](https://community.drivendata.org/t/using-pre-trained-models/6178/1 "2021-06-10T10:02:22Z")

</div>

Given prohibition of external data, question is whether pre-trained model weights that are publicly available under permissive license like MIT or Apache can be used in solutions?

---

<div class="post-metadata">

**Author:** ![glipstein](https://yyz2.discourse-cdn.com/flex028/user_avatar/community.drivendata.org/glipstein/32/920_2.png) [@glipstein](https://community.drivendata.org/u/glipstein)\
**Post date:** [June 10, 2021, 4:30pm UTC](https://community.drivendata.org/t/using-pre-trained-models/6178/2 "2021-06-10T16:30:09Z")

</div>

Hi @tetelias - Thanks for asking about this. Here is the note on pre-trained models from the [Problem Description](https://www.drivendata.org/competitions/78/overhead-geopose-challenge/page/372/).

> External data is not allowed in this competition. Participants can use pre-trained computer vision models as long as they were available freely and openly in that form at the start of the competition.

Hope that helps!