Skip to content

iSTFTNet : Fast and Lightweight Mel-spectrogram Vocoder Incorporating Inverse Short-time Fourier Transform

License

Notifications You must be signed in to change notification settings

rishikksh20/iSTFTNet-pytorch

Repository files navigation

iSTFTNet : Fast and Lightweight Mel-spectrogram Vocoder Incorporating Inverse Short-time Fourier Transform

This repo try to implement iSTFTNet : Fast and Lightweight Mel-spectrogram Vocoder Incorporating Inverse Short-time Fourier Transform specifically model C8C8I. Disclaimer : This repo is build for testing purpose. The code is not optimized for performance.

Training :

python train.py --config config_v1.json

Note:

  • We are able to get good quality of audio with 30 % less training compared to original hifigan.
  • This model approx 60 % faster than counterpart hifigan.

Citations :

@inproceedings{kaneko2022istftnet,
title={{iSTFTNet}: Fast and Lightweight Mel-Spectrogram Vocoder Incorporating Inverse Short-Time Fourier Transform},
author={Takuhiro Kaneko and Kou Tanaka and Hirokazu Kameoka and Shogo Seki},
booktitle={ICASSP},
year={2022},
}

References:

About

iSTFTNet : Fast and Lightweight Mel-spectrogram Vocoder Incorporating Inverse Short-time Fourier Transform

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages