# Token\_type\_ids and DistilBert

**URL:** <https://discuss.huggingface.co/t/token-type-ids-and-distilbert/21172>\
**Category:** Models\
**Created:** [August 2, 2022, 9:11pm UTC](https://discuss.huggingface.co/t/token-type-ids-and-distilbert/21172 "2022-08-02T21:11:39Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![amaiya](https://avatars.discourse-cdn.com/v4/letter/a/8797f3/32.png) [@amaiya](https://discuss.huggingface.co/u/amaiya)\
**Post date:** [August 2, 2022, 9:11pm UTC](https://discuss.huggingface.co/t/token-type-ids-and-distilbert/21172/1 "2022-08-02T21:11:39Z")

</div>

Some models accept `token_type_ids` and others do not. `TFDistilBertForSequenceClassification` does **not**. But, the `call` method used to accept a `**kwargs` parameter. This allowed users to always supply `token_type_ids` to any model and if the model did not need it it would just be ignored.

Since 4.18.0, the ` **kwargs` argument seems to have been removed. This means we have to keep track of which models accept the `token_type_ids` argument and ensure that `token_type_ids` is** not **sent to the `call` method if it is a model like DistilBert, correct? To me it seems like having the `** kwargs` argument was a good thing, so I was wondering if there was something obvious I was missing.