# Stopping Criteria List - OPT model

**URL:** <https://discuss.huggingface.co/t/stopping-criteria-list-opt-model/20677>\
**Category:** Beginners\
**Created:** [July 21, 2022, 9:46am UTC](https://discuss.huggingface.co/t/stopping-criteria-list-opt-model/20677 "2022-07-21T09:46:13Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![anujn](https://avatars.discourse-cdn.com/v4/letter/a/5f9b8f/32.png) [@anujn](https://discuss.huggingface.co/u/anujn)\
**Post date:** [July 21, 2022, 9:46am UTC](https://discuss.huggingface.co/t/stopping-criteria-list-opt-model/20677/1 "2022-07-21T09:46:13Z")

</div>

Dear HF,

I am trying to implement the stopping criteria list method on model.generate.

The docs say to use the BERT tokeniser to generate the input ids, and the scores. However this is not the tokenizer associated with the OPT model.

Should I be generating them with the OPT tokenizer or the BERT one?

---

<div class="post-metadata">

**Author:** ![Lyma](https://avatars.discourse-cdn.com/v4/letter/l/e68b1a/32.png) [@Lyma](https://discuss.huggingface.co/u/Lyma)\
**Post date:** [October 14, 2022, 2:59pm UTC](https://discuss.huggingface.co/t/stopping-criteria-list-opt-model/20677/2 "2022-10-14T14:59:29Z")

</div>

Hi!

Please, check my reply in the other question of yours: [Implimentation of Stopping Criteria List - #2 by Lyma](https://discuss.huggingface.co/t/implimentation-of-stopping-criteria-list/20040/2)
