rishikksh20 / TalkNet2-pytorch

TalkNet 2: Non-Autoregressive Depth-Wise Separable Convolutional Model for Speech Synthesis with Explicit Pitch and Duration Prediction.
MIT License
85 stars 6 forks source link
pytorch speech speech-synthesis text-to-speech tts

TalkNet 2 [WIP]

TalkNet 2: Non-Autoregressive Depth-Wise Separable Convolutional Model for Speech Synthesis with Explicit Pitch and Duration Prediction.


Official TalkNet 2 repo here

Work remains:

Citation:

@misc{beliaev2021talknet,
      title={TalkNet 2: Non-Autoregressive Depth-Wise Separable Convolutional Model Stanislav Beliaev, Boris Ginsburgfor Speech Synthesis with Explicit Pitch and Duration Prediction}, 
      author={Stanislav Beliaev and Boris Ginsburg},
      year={2021},
      eprint={2104.08189},
      archivePrefix={arXiv},
      primaryClass={eess.AS}
}