# Train a t5 model

**URL:** </static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Ftrain-a-t5-model%2F53425%3E%5C
**Category:** 🤗Transformers\
**Created:** [September 4, 2023, 6:07am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Ftrain-a-t5-model%2F53425 "2023-09-04T06:07:53Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![sanjay782](https://avatars.discourse-cdn.com/v4/letter/s/dbc845/32.png) [@sanjay782](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fsanjay782)%5C
**Post date:** [September 4, 2023, 6:07am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Ftrain-a-t5-model%2F53425%2F1 "2023-09-04T06:07:53Z")

</div>

I want to pre train a t5 model for question answer task with own dataset.The dataset contain 10 question answer for each paragraph.How to give input to the t5 model.

---

<div class="post-metadata">

**Author:** ![vipin-rv2307](https://avatars.discourse-cdn.com/v4/letter/v/839c29/32.png) [@vipin-rv2307](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fvipin-rv2307)%5C
**Post date:** [September 4, 2023, 10:48am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Ftrain-a-t5-model%2F53425%2F2 "2023-09-04T10:48:39Z")

</div>

If you have 10 question and answer pairs for each context, Your input structure should be  
source\_text-\>context  
target\_text-\>QA pairs separated by any tokens , for example → [SEP] like that