# Should we save the tokenizer state over domain adaptation?

**URL:** </static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fshould-we-save-the-tokenizer-state-over-domain-adaptation%2F13324%3E%5C
**Category:** Beginners\
**Created:** [January 3, 2022, 3:28pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fshould-we-save-the-tokenizer-state-over-domain-adaptation%2F13324 "2022-01-03T15:28:08Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![MahdiA](https://sea2.discourse-cdn.com/hellohellohello/user_avatar/discuss.huggingface.co/mahdia/32/2700_2.png) [@MahdiA](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2FMahdiA)%5C
**Post date:** [January 3, 2022, 3:28pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fshould-we-save-the-tokenizer-state-over-domain-adaptation%2F13324%2F1 "2022-01-03T15:28:08Z")

</div>

I am going to do domain adaptation over my dataset with BERT. When I train the BERT model, should I save the tokenizer state? I mean, will tokenizer state change over training the model?

---

<div class="post-metadata">

**Author:** ![sgugger](https://sea2.discourse-cdn.com/hellohellohello/user_avatar/discuss.huggingface.co/sgugger/32/2291_2.png) [@sgugger](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fsgugger)%5C
**Post date:** [January 10, 2022, 3:49pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fshould-we-save-the-tokenizer-state-over-domain-adaptation%2F13324%2F2 "2022-01-10T15:49:18Z")

</div>

No, the tokenizer is not changed by the model fine-tuning on a new dataset.