# PubMedQA, Preprocessing

**URL:** <https://discuss.huggingface.co/t/pubmedqa-preprocessing/48670>\
**Category:** 🤗Transformers\
**Created:** [July 30, 2023, 2:37am UTC](https://discuss.huggingface.co/t/pubmedqa-preprocessing/48670 "2023-07-30T02:37:58Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![tyutfyu](https://avatars.discourse-cdn.com/v4/letter/t/58956e/32.png) [@tyutfyu](https://discuss.huggingface.co/u/tyutfyu)\
**Post date:** [July 30, 2023, 2:37am UTC](https://discuss.huggingface.co/t/pubmedqa-preprocessing/48670/1 "2023-07-30T02:37:58Z")

</div>

Hi,  
I am trying to use PubMED QA dataset, when I am using long\_answer and try to preprocessing the data  
most of end/stop position the answers are taking 0 as position and that is because the long-answer are not necessarily located in the context  
any thoughts about how that can be solved  
thanks