FakeYou SoftVC
==============

This work is based off research by Michael and Paul, leveraging bshall's SoftVC:

https://bshall.github.io/soft-vc/

These are the repos contained herein: 

* https://github.com/bshall/hubert
* https://github.com/bshall/acoustic-model
* https://github.com/bshall/hifigan

And this is our initial notebook (from Vegito): 

* [SoftVCTraceAndOnnx.ipynb](https://colab.research.google.com/drive/1g0YsQQY8fzXHmZUWAMzQO60c6OUlyKBl#scrollTo=Sf_rjYw0P83I)


From Paul: 

* [softvc_pynb.ipynb](https://drive.google.com/file/d/1FzpL3Ze3X9ESt_nB4w9rTgnfGJxunha0/view?usp=share_link)
* [softvc_hifigan_fine_tuning.ipynb](https://drive.google.com/file/d/1X6Z5sLogpzTx8qEdbQ5Ub3N4U8_ZikPy/view?usp=share_link)
* [SoftVCTraceAndOnnx.ipynb](https://drive.google.com/file/d/13n-rFTqReBGJBQMB7N9nQClySZw8KzLV/view?usp=share_link)