# yl4579/StyleTTS2

StyleTTS 2: Towards Human-Level Text-to-Speech through Style Diffusion and Adversarial Training with Large Speech Language Models

Repository: https://github.com/yl4579/StyleTTS2
Canonical: https://ross.abutalabs.com/products/styletts2
Language: Python
License: MIT
License Family: permissive
Topics: deep-learning, pytorch, speaker-adaptation, speech-synthesis, text-to-speech, tts, wavlm, diffusion-models, latent-diffusion, latent-diffusion-models, adversarial-training, gan
Last push: 2024-08-10T00:48:18+00:00

## Health v2 (maintenance only)
Score: 29/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 84
- inputs: {"age_days": 1177, "days_push": 754, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 6336, forks 700 (observed 2026-08-28T04:09:41.817316+00:00)

## What it is
StyleTTS 2 is a PyTorch text-to-speech model that uses style diffusion and adversarial training with large speech language models (e.g., WavLM) to achieve human-level speech synthesis. It supports single-speaker and multi-speaker synthesis, zero-shot speaker adaptation, and finetuning for new speakers.

## Use cases
- convert text to natural-sounding speech
- clone a voice from a short reference audio clip
- synthesize speech without a reference recording using style diffusion
- finetune a TTS model on a new speaker's voice
- run zero-shot multi-speaker speech synthesis
- generate audiobook-style narration from text

## When to choose
- you need state-of-the-art, human-level TTS quality in English
- you want zero-shot speaker adaptation from pretrained models
- you need a research-grade TTS model you can finetune
- you want speech synthesis without requiring reference audio

## When to avoid
- you need production TTS for many languages (primarily English datasets)
- you lack a GPU or want lightweight on-device inference
- you need a simple plug-and-play API rather than research code
- you need non-Python integration out of the box

## Facets
- artifact type: library
- maturity: active
- function: tts, machine-learning, deep-learning
- domain: speech-processing, deep-learning, machine-learning
- platform: python, cross-platform
- tags: text-to-speech, speech-synthesis, diffusion-models, adversarial-training, pytorch, zero-shot-voice-cloning, wavlm, style-diffusion, gpu

## Member repositories
- yl4579/StyleTTS2 (main) score 29

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:41.817316+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:46:09.182147+00:00, confidence not recorded.
  - readme: https://github.com/yl4579/StyleTTS2 (fetched 2026-08-28T04:09:41.817316+00:00, sha d270c4f5af41)
- Data as of 2026-08-30T08:39:29.467469+00:00.