# Effects on other models in pipe with training NER

**URL:** <https://support.prodi.gy/t/effects-on-other-models-in-pipe-with-training-ner/554>\
**Category:** Uncategorized\
**Tags:** ner, spacy, solved\
**Created:** [May 21, 2018, 7:21pm UTC](https://support.prodi.gy/t/effects-on-other-models-in-pipe-with-training-ner/554 "2018-05-21T19:21:55Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![mhigginslp](https://avatars.discourse-cdn.com/v4/letter/m/839c29/32.png) [@mhigginslp](https://support.prodi.gy/u/mhigginslp)\
**Post date:** [May 21, 2018, 7:21pm UTC](https://support.prodi.gy/t/effects-on-other-models-in-pipe-with-training-ner/554/1 "2018-05-21T19:21:55Z")

</div>

Subject pretty much says it all. Using default en\_core\_web\_lg or sm what other elements are effected by the training of the NER?

---

<div class="post-metadata">

**Author:** ![honnibal](https://sea2.discourse-cdn.com/flex020/user_avatar/support.prodi.gy/honnibal/32/35_2.png) [@honnibal](https://support.prodi.gy/u/honnibal)\
**Post date:** [May 21, 2018, 7:55pm UTC](https://support.prodi.gy/t/effects-on-other-models-in-pipe-with-training-ner/554/2 "2018-05-21T19:55:34Z")

</div>

Nothing should be affected. The NER doesn’t share any trainable weights with the parser or tagger. The NER also doesn’t backprop into the vectors used for the similarity calculations.