# MQ\_baseline differences

**URL:** <https://discuss.ego4d-data.org/t/mq-baseline-differences/312>\
**Category:** Ego4D Challenges\
**Tags:** moments\
**Created:** [May 8, 2023, 9:30am UTC](https://discuss.ego4d-data.org/t/mq-baseline-differences/312 "2023-05-08T09:30:31Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![fedegonzal](https://yyz1.discourse-cdn.com/flex007/user_avatar/discuss.ego4d-data.org/fedegonzal/32/54_2.png) [@fedegonzal](https://discuss.ego4d-data.org/u/fedegonzal)\
**Post date:** [May 8, 2023, 9:30am UTC](https://discuss.ego4d-data.org/t/mq-baseline-differences/312/1 "2023-05-08T09:30:31Z")

</div>

Hello, I’m confused with MQ\_baseline published results. There is two MQ\_baseline on Leaderboard, the first one with Recall 44.24 and average\_mAP 23.99, but the other has 24.25 and 5.68

As I understand, MQ\_baseline is the official baseline published. Why it’s twice?

Also I clonned the official Github Repo and trained the official model using slowfast features… getting similar results to 24.25 and 5.68 (the worst MQ\_baseline). So… which setup has the “better” MQ\_baseline? it’s the same model, the same repo? I’d like to reproduce it correctly.

Thanks!

---

<div class="post-metadata">

**Author:** ![srama2512](https://avatars.discourse-cdn.com/v4/letter/s/eb9ed0/32.png) [@srama2512](https://discuss.ego4d-data.org/u/srama2512)\
**Post date:** [May 12, 2023, 8:38pm UTC](https://discuss.ego4d-data.org/t/mq-baseline-differences/312/2 "2023-05-12T20:38:17Z")

</div>

@fedegonzal - Thanks for your interest in the MQ challenge. The MQ\_baseline is based on the last year’s winning entry. We have updated the details on the baseline here (includes pointers to the codebase): [https://eval.ai/web/challenges/challenge-page/1626/overview](https://eval.ai/web/challenges/challenge-page/1626/overview)