# Random seed for weight initialization and data order

**URL:** </static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Frandom-seed-for-weight-initialization-and-data-order%2F14989%3E%5C
**Category:** 🤗Transformers\
**Created:** [February 21, 2022, 6:23pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Frandom-seed-for-weight-initialization-and-data-order%2F14989 "2022-02-21T18:23:42Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![phosseini](https://sea2.discourse-cdn.com/hellohellohello/user_avatar/discuss.huggingface.co/phosseini/32/26690_2.png) [@phosseini](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fphosseini)%5C
**Post date:** [February 21, 2022, 6:23pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Frandom-seed-for-weight-initialization-and-data-order%2F14989%2F1 "2022-02-21T18:23:42Z")

</div>

A simple question, I wonder if the `seed` argument in the `TrainingArguments ` will be used for both weight initialization of a model and shuffling the data? And if it’s used only for weight initialization, what would be a good practice for data shuffling? Just randomly shuffle it ourselves?