# \[Question\] Wav2vec2 word times

**URL:** </static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fquestion-wav2vec2-word-times%2F5039%3E%5C
**Category:** 🤗Transformers\
**Created:** [March 24, 2021, 7:58am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fquestion-wav2vec2-word-times%2F5039 "2021-03-24T07:58:41Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![irux](https://avatars.discourse-cdn.com/v4/letter/i/9fc29f/32.png) [@irux](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Firux)%5C
**Post date:** [March 24, 2021, 7:58am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fquestion-wav2vec2-word-times%2F5039%2F1 "2021-03-24T07:58:41Z")

</div>

Hello, did someone already try to get the timestamps of the words while decoding the audio with wav2vec2 model?

Thank you!

---

<div class="post-metadata">

**Author:** ![KB-hugging](https://avatars.discourse-cdn.com/v4/letter/k/a5b964/32.png) [@KB-hugging](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2FKB-hugging)%5C
**Post date:** [June 24, 2021, 1:55pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fquestion-wav2vec2-word-times%2F5039%2F2 "2021-06-24T13:55:50Z")

</div>

I have created some hacky code to do so and have posted it in this [huggingface github issue](https://github.com/huggingface/transformers/issues/11307). You can find it here: [Getting time offsets of beginning and end of each word in Wav2Vec2 · Issue #11307 · huggingface/transformers · GitHub](https://github.com/huggingface/transformers/issues/11307#issuecomment-867648870)

---

<div class="post-metadata">

**Author:** ![jonaskratochvil](https://avatars.discourse-cdn.com/v4/letter/j/b9bd4f/32.png) [@jonaskratochvil](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fjonaskratochvil)%5C
**Post date:** [June 24, 2021, 7:01pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fquestion-wav2vec2-word-times%2F5039%2F3 "2021-06-24T19:01:59Z")

</div>

Hi, I have successfully used this repository: [GitHub - lumaku/ctc-segmentation: Segment an audio file and obtain utterance alignments. (Python package)](https://github.com/lumaku/ctc-segmentation) to obtain the word level timestamps from wav2vec.