# Load NLU and Core separately in single service

**URL:** <https://forum.rasa.com/t/load-nlu-and-core-separately-in-single-service/28704>\
**Category:** Rasa Open Source\
**Created:** [May 11, 2020, 4:41pm UTC](https://forum.rasa.com/t/load-nlu-and-core-separately-in-single-service/28704 "2020-05-11T16:41:09Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![sjtilney](https://avatars.discourse-cdn.com/v4/letter/s/f17d59/32.png) [@sjtilney](https://forum.rasa.com/u/sjtilney)\
**Post date:** [May 11, 2020, 4:41pm UTC](https://forum.rasa.com/t/load-nlu-and-core-separately-in-single-service/28704/1 "2020-05-11T16:41:09Z")

</div>

My training data set has become large, and training is now taking over 1 hour. Training time is now having a major impact of developer productivity. In order to reduce training time, I would like to train my NLU model and Core model separately and in parallel. While Rasa allows me to train separately, I don’t see a way to load them separately without having two separate services running. Ideally I would like to do something like this

$ rasa run --core \<core\_model\> --nlu \<nlu\_model\>

Are there any ways to achieve this?