# Sentences' embeddings from BERT cross-encoder

**URL:** <https://discuss.huggingface.co/t/sentences-embeddings-from-bert-cross-encoder/28313>\
**Category:** 🤗Transformers\
**Created:** [December 22, 2022, 8:26pm UTC](https://discuss.huggingface.co/t/sentences-embeddings-from-bert-cross-encoder/28313 "2022-12-22T20:26:46Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![celsofranssa](https://sea2.discourse-cdn.com/hellohellohello/user_avatar/discuss.huggingface.co/celsofranssa/32/27854_2.png) [@celsofranssa](https://discuss.huggingface.co/u/celsofranssa)\
**Post date:** [December 22, 2022, 8:26pm UTC](https://discuss.huggingface.co/t/sentences-embeddings-from-bert-cross-encoder/28313/1 "2022-12-22T20:26:46Z")

</div>

Hello,

Does it make sense to forward pass a sentence pair to BERT to get the maximum iterations between all tokens and subsequently get the vectors corresponding to each sentence?