Back: Catalog

StyleTTS2

StyleTTS 2 is a TTS model that synthesizes from text to sound using adversarial training with style diffusion and large speech language models. Their style models as a hidden random variable and allows for the production of text-appropriate style without the need for reference speech. It uses large pre-trained models such as WavLM as distinctive and achieves human-level synthesis in single and multi-speaker data sets.

ST

StyleTTS2

55

yl4579 · Voice & Text-to-Speech

STARS6.4K
FORKS712
COMMITS112
CONTRIBUTORS12
VERSION—
UPDATED2 yr ago
PythonDiffusionİnsan SeviyeDuygusalMIT

Common commercial products in this category: ElevenLabs, Amazon Polly, Google Cloud Text-to-Speech

What can it do?
  • You train and infer on single-speaker LJSpeech or multi-speaker VCTK and LibriTTS datasets
  • Using the provided base pre-trained models, you run the fine-tuning script for new speakers
  • You take advantage of diffusion models that produce the most appropriate style for text without the need for reference speech
Heads up
  • • Last update 2 years ago
User reviews
  • • NeutralHacker News

    StyleTTS2 – open-source Eleven-Labs-quality Text To Speech

    “So, we've got this open-source TTS wizardry going on, which is kinda like if Siri had a caffeine overdose - faster, snappier, and way more fun at parties. This thing is running on gaming rigs with beefy GPUs, and it's apparently so user-friendly, even your grandma could set it…”

  • • NeutralHacker News

    StyleTTS2 – open-source Eleven-Labs-quality Text To Speech

    “This is great! Nice work. I made my own whisper & auto-typer which types what you say (forked whisper-typer). I added OpenAI Q/A and RAG query feature so I could ask it questions (instead of auto keystroke typing) by voice command. For responses to questions, I used Eleven Labs…”

GitHub

My note

Alternative pages

ElevenLabs alternativesAmazon Polly alternativesGoogle Cloud Text-to-Speech alternatives

More in: Voice & Text-to-Speech