# Where to find documentation on dataset format for finetuning

**URL:** <https://discuss.huggingface.co/t/where-to-find-documentation-on-dataset-format-for-finetuning/57741>\
**Category:** Beginners\
**Created:** [October 7, 2023, 6:34pm UTC](https://discuss.huggingface.co/t/where-to-find-documentation-on-dataset-format-for-finetuning/57741 "2023-10-07T18:34:13Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![pvelosipednikov](https://avatars.discourse-cdn.com/v4/letter/p/90db22/32.png) [@pvelosipednikov](https://discuss.huggingface.co/u/pvelosipednikov)\
**Post date:** [October 7, 2023, 6:34pm UTC](https://discuss.huggingface.co/t/where-to-find-documentation-on-dataset-format-for-finetuning/57741/1 "2023-10-07T18:34:13Z")

</div>

I want to finetune a Bart model with an AutoModelForSeq2SeqLM head.

Where do I find the format to preprocess the dataset into?  
I think the answer may be the arguments in the [forward method](https://huggingface.co/docs/transformers/model_doc/bart#transformers.BartModel.forward) of the tokenizer?

So in my case, I would need to make sure that my dataset has the following columns:  
input\_ids, attention\_mask, decoder\_input\_ids, decoder\_attention\_mask  
?

Would this potentially change depending on what head I use for finetuning, and if so where would I turn to see those requirements?