# How to determine the data format when creating a custom dataset for a given task?

**URL:** <https://discuss.huggingface.co/t/how-to-determine-the-data-format-when-creating-a-custom-dataset-for-a-given-task/36952>\
**Category:** 🤗Transformers\
**Created:** [April 18, 2023, 8:19am UTC](https://discuss.huggingface.co/t/how-to-determine-the-data-format-when-creating-a-custom-dataset-for-a-given-task/36952 "2023-04-18T08:19:32Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![jimothyhalpert7](https://avatars.discourse-cdn.com/v4/letter/j/f19dbf/32.png) [@jimothyhalpert7](https://discuss.huggingface.co/u/jimothyhalpert7)\
**Post date:** [April 18, 2023, 8:19am UTC](https://discuss.huggingface.co/t/how-to-determine-the-data-format-when-creating-a-custom-dataset-for-a-given-task/36952/1 "2023-04-18T08:19:32Z")

</div>

Trying to fine tune a DialoGPT model on the cornell dataset.  
It needs to be preprocessed.

Where do I find the format that the model expects the data to be in? For this, and for other models?