# Finetune / retrain wav2vec2 encoder in self-supervised manner

**URL:** <https://discuss.huggingface.co/t/finetune-retrain-wav2vec2-encoder-in-self-supervised-manner/15483>\
**Category:** Beginners\
**Created:** [March 8, 2022, 1:11am UTC](https://discuss.huggingface.co/t/finetune-retrain-wav2vec2-encoder-in-self-supervised-manner/15483 "2022-03-08T01:11:53Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![Kuray107](https://avatars.discourse-cdn.com/v4/letter/k/a87d85/32.png) [@Kuray107](https://discuss.huggingface.co/u/Kuray107)\
**Post date:** [March 8, 2022, 1:11am UTC](https://discuss.huggingface.co/t/finetune-retrain-wav2vec2-encoder-in-self-supervised-manner/15483/1 "2022-03-08T01:11:53Z")

</div>

Hi everyone, is there any member function in the wav2vec2ForCTC model that allows us to use unlabeled audio data to fine-tune / retrain the wav2vec2 encoder? I think we should be able to do this since there are so many public self-pretrained models (e.x. facebook/wav2vec2-large-lv60), but I can’t find any related information neither on this forum nor the official documentation.