comprehensive-transformer-tts - v0.2.1

Fix and update codebase & pre-trained models with demo samples 1. Fix variance adaptor to make it work with all combinations of building block and variance type/level 2. Update pre-trained models with demo samples of LJSpeech and VCTK under "transformerfs2" building block and "cwt" pitch conditioning 3. Share the result of ablation studies of comparing "transformer" vs. "transformerfs2" paired among three types of pitch conditioning ("frame", "ph", and "cwt")

- Python
Published by keonlee9420 about 4 years ago

comprehensive-transformer-tts - v0.2.0

A lot of improvements with new features! 1) Prepare two different types of data pipeline in preprocessor to maximize unsupervised/supervised duration modelings 2) Adopt wavelet for pitch modeling & loss 3) Add fine-trained duration loss 4) Apply var_start_steps for better model convergence, especially under unsupervised duration modeling 5) Remove dependency of energy modeling on pitch variance 6) Add "transformerfs2" building block, which is more close to the original FastSpeech2 paper 7) Add two types of prosody modeling methods 8) Loss camparison on validation set: - LJSpeech - blue: v0.1.1 / green: v0.2.0

<img src="./img/losscomparison_ljs.png" width="80%">

- VCTK - skyblue: v0.1.1 / orange: v0.2.0
<p align="center">
    <img src="./img/loss_comparison_vctk.png" width="80%">
</p>

- Python
Published by keonlee9420 over 4 years ago

comprehensive-transformer-tts - v0.1.1

- Python
Published by keonlee9420 over 4 years ago

comprehensive-transformer-tts - v0.1.0

- Python
Published by keonlee9420 over 4 years ago