# Use transformer without position embeddings being added?

**URL:** <https://discuss.huggingface.co/t/use-transformer-without-position-embeddings-being-added/6779>\
**Category:** Beginners\
**Created:** [June 13, 2021, 2:45pm UTC](https://discuss.huggingface.co/t/use-transformer-without-position-embeddings-being-added/6779 "2021-06-13T14:45:03Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![abpalmarini](https://avatars.discourse-cdn.com/v4/letter/a/bcef8e/32.png) [@abpalmarini](https://discuss.huggingface.co/u/abpalmarini)\
**Post date:** [June 13, 2021, 2:45pm UTC](https://discuss.huggingface.co/t/use-transformer-without-position-embeddings-being-added/6779/1 "2021-06-13T14:45:03Z")

</div>

Hi,

I can’t seem to find any option to disable position embeddings being added in models like BERT and GPT2. I’d like to use such a model for a non NLP task where position is irrelevant. Is it possible to do this in Hugging Face?

Thanks