Skip to content
SPECIFICTION
Get listed

StyleTTS 2

styletts2.github.io

WindowsLinuxSelf-hosted
Visit styletts2.github.io
Spec score 7.5 #163 of 233 in Text-to-speech tools

Spec sheet

Purpose
StyleTTS 2 is a text-to-speech model that uses style diffusion and adversarial training with large speech language models.Read 13 Oct 2026 · styletts2.github.io
Platforms
windows,linux,self_hostedRead 20 Sep 2026 · styletts2.github.io
Code license
The repository states that its code is licensed under the MIT License.Read 13 Oct 2026 · github.com
Compute requirements
The repository notes that training or inference can run into out-of-memory issues and recommends reducing batch size or maximum audio length; it also says older GPUs may produce background noise.Read 13 Oct 2026 · github.com
Dependencies
The repository says its inference depends on a GPL-licensed package that is not included directly, and also links to a fully MIT-licensed package using gruut with lower output quality due to a phonemizer mismatch.Read 13 Oct 2026 · github.com
Evaluation
The makers report that StyleTTS 2 surpasses human recordings on single-speaker LJSpeech and matches them on multispeaker VCTK, as judged by native English speakers.Read 12 Oct 2026 · styletts2.github.io
Expressiveness
The sample page demonstrates expressive speech, style transfer, longform narration, and speech diversity.Read 13 Oct 2026 · styletts2.github.io
Inference
The repository provides inference notebooks and pretrained model downloads for LJSpeech and LibriTTS.Read 13 Oct 2026 · github.com
Installation
The repository lists Python 3.7 or newer as a prerequisite and gives additional installation instructions for Windows.Read 13 Oct 2026 · github.com
Known limitation
The repository says the distributed-data-parallel version of second-stage training does not work, so the current version uses data parallelism.Read 13 Oct 2026 · github.com
Language support
The repository says non-English training requires a pretrained PL-BERT model for the language and points to a multilingual PL-BERT supporting 14 languages.Read 13 Oct 2026 · github.com
Limitations
The README says the second-stage training code uses data parallelism because its distributed data parallel version does not work.Read 12 Oct 2026 · github.com
Linux dependency
The demo instructions say to install phonemizer and espeak-ng; the example espeak-ng command uses apt-get.Read 12 Oct 2026 · github.com
Model use
For pretrained models, the repository asks users to disclose synthesized speech unless they have permission to use the cloned voice, and says public voice use requires speaker permission or disclosure.Read 13 Oct 2026 · github.com
Online demo
The repository links to a Hugging Face online demo.Read 13 Oct 2026 · github.com
Open source license
The repository code is licensed under the MIT License.Read 12 Oct 2026 · github.com
Pretrained models
The repository links pretrained models for LJSpeech and LibriTTS.Read 12 Oct 2026 · github.com
Speaker adaptation
The project reports that its LibriTTS-trained model outperforms previous publicly available models for zero-shot speaker adaptation.Read 13 Oct 2026 · styletts2.github.io
Speech naturalness
The model uses pre-trained speech language models such as WavLM as discriminators and differentiable duration modeling to improve speech naturalness.Read 12 Oct 2026 · styletts2.github.io
Speech quality
The authors report that it surpasses human recordings on the single-speaker LJSpeech dataset and matches them on multispeaker VCTK, as judged by native English speakers.Read 13 Oct 2026 · styletts2.github.io
Style generation
It models speaking styles as latent random variables and uses diffusion to generate a suitable style without reference speech.Read 13 Oct 2026 · styletts2.github.io
Style transfer
The model can transfer an emotional style sampled from emotional text to input text that has no emotional content.Read 12 Oct 2026 · styletts2.github.io
Voice cloning
YesRead 20 Sep 2026 · styletts2.github.io
Voice use terms
For pretrained models, the README says users must have permission to clone a speaker's voice or publicly disclose that the generated voice is synthesized.Read 12 Oct 2026 · github.com
Windows support
The setup instructions provide an additional PyTorch installation command for Windows.Read 12 Oct 2026 · github.com

Where Specifiction places it

Alternatives to StyleTTS 2All alternatives →