OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Audio & Voice AI

Glow Tts

Glow Tts is an open-source, flow-based text-to-speech model implemented in Python with PyTorch. It generates speech representations in parallel and finds monotonic text-to-speech alignments through dynamic programming, without an external aligner. Its repository provides training and inference examples, pretrained model links, and an example of connecting inference to HiFi-GAN for audio output.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Glow Tts — OpenViST project profile in Audio & Voice AI.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Glow Tts when you need a repository-based model for experimenting with or building a self-hosted text-to-speech pipeline, and can manage its Python/PyTorch environment and model dependencies.

Who it is for

  • Developers building self-hosted text-to-speech applications.
  • Researchers experimenting with speech synthesis, model training or multi-speaker generation.

Documented uses

  • Generate speech representations from text through model inference.
  • Train Glow-TTS using the documented LJ Speech dataset setup.
  • Explore multi-speaker speech generation.
  • Connect inference to HiFi-GAN to produce audio from synthesized representations.
Before you choose

Requirements & limits

The repository is public and specifies an MIT license.

  • Training setup involves LJ Speech, WaveGlow and building the Cython alignment code.
  • The README's example environment uses older, explicitly listed dependency versions.
  • The repository provides inference notebooks and pretrained model links, while audio output can use the documented HiFi-GAN path.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Prepare the documented training setup

Set up the dataset and required components for the repository's Glow-TTS training example.

Before you start

  • Python 3.6.9, PyTorch 1.2.0, Cython 0.29.12, librosa 0.7.1, NumPy 1.16.4 and SciPy 1.3.0, as listed in the README.
  • The LJ Speech dataset.
  • The repository's WaveGlow submodule and a pretrained WaveGlow model.

Steps

  1. Download and extract LJ Speech, then rename or link its wavs folder as DUMMY.

  2. Initialize and update the WaveGlow submodule.

  3. Place the pretrained WaveGlow model in the waveglow folder.

  4. Build the Monotonic Alignment Search code from the monotonic_align directory using the documented Cython build command.

Expected outcome

The README's documented data and component preparation is complete for the training example.

Guide limitations

  • The README separately provides a training command but this guide covers setup only.
  • The listed environment versions describe the environment used by the project; they are not a compatibility guarantee.
Documentation 1 ↗
COMPARE OPTIONS

Compare options for Model inference and serving

These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.

ProjectDocumented purposePlatformAccess and requirements
Glow TtsGenerate speech representations from text through model inference.Linux, Windows, macOS, Self-hostedThe repository is public and specifies an MIT license.
Speech Denoising WavenetRun a pretrained model on noisy speech.Linux, CLIThe repository is public and identifies its license as MIT.
WavtokenizerEncode audio into discrete codes for downstream audio-language-model work.Linux, Windows, macOSThe repository is public and lists an MIT license. The README links to model checkpoints and collections on Hugging Face.
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
7 November 2026
Original website title
jaywalnut310/glow-tts · Official homepage title · retained in its source language
How we review sourcesSuggest a correction