Glow TTS¶
Glow TTS is a normalizing flow model for text-to-speech. It is built on the generic Glow model that is previously used in computer vision and vocoder models. It uses “monotonic alignment search” (MAS) to fine the text-to-speech alignment and uses the output to train a separate duration predictor network for faster inference run-time.
Important resources & papers¶
GlowTTS: https://arxiv.org/abs/2005.11129
Glow (Generative Flow with invertible 1x1 Convolutions): https://arxiv.org/abs/1807.03039
Normalizing Flows: https://blog.evjang.com/2018/01/nf1.html