# Multilingual NER Extraction

**URL:** </static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fmultilingual-ner-extraction%2F20303%3E%5C
**Category:** Models\
**Created:** [July 13, 2022, 1:08am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fmultilingual-ner-extraction%2F20303 "2022-07-13T01:08:22Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![trwa4172](https://avatars.discourse-cdn.com/v4/letter/t/e95f7d/32.png) [@trwa4172](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Ftrwa4172)%5C
**Post date:** [July 13, 2022, 1:08am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fmultilingual-ner-extraction%2F20303%2F1 "2022-07-13T01:08:22Z")

</div>

Hi Everyone!

I had a really quick question about approaching multilingual NER extraction for domain specific use cases (tracking shipments/making quotes). When it comes to transfer learning on multilingual models for NER, would it be better to take a model like [bert-base-multilingual-cased](https://huggingface.co/bert-base-multilingual-cased) and fine-tune it for domain specific NER, or take a more defined model like [bert-base-multilingual-cased-ner-hrl](https://huggingface.co/Davlan/bert-base-multilingual-cased-ner-hrl?text=Mein+Name+ist+Trevor+und+ich+lebe+in+Berlin) and train it for domain specific words?

---

<div class="post-metadata">

**Author:** ![shreyans92dhankhar](https://sea2.discourse-cdn.com/hellohellohello/user_avatar/discuss.huggingface.co/shreyans92dhankhar/32/4909_2.png) [@shreyans92dhankhar](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fshreyans92dhankhar)%5C
**Post date:** [July 13, 2022, 7:53am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fmultilingual-ner-extraction%2F20303%2F2 "2022-07-13T07:53:01Z")

</div>

I also had a domain specific use case for NER and i followed first approach which worked well.