# Wav2vec - \<s\>\</s\> tokens

**URL:** <https://discuss.huggingface.co/t/wav2vec-s-s-tokens/13872>\
**Category:** Models\
**Created:** [January 18, 2022, 8:07pm UTC](https://discuss.huggingface.co/t/wav2vec-s-s-tokens/13872 "2022-01-18T20:07:45Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![bhemmat](https://avatars.discourse-cdn.com/v4/letter/b/13edae/32.png) [@bhemmat](https://discuss.huggingface.co/u/bhemmat)\
**Post date:** [January 18, 2022, 8:07pm UTC](https://discuss.huggingface.co/t/wav2vec-s-s-tokens/13872/1 "2022-01-18T20:07:45Z")

</div>

The wav2vec2.0 base 960h model never seems to return a beginning of sentence or end of sentence token (or ’ or unknown, so far)–using greedy decoding. Is that expected? I can’t seem to find this discussed anywhere. Or am I just feeding in audio that is too difficult for the model to determine the eos/bos? If so, can someone provide a counter-example?