# lifeiteng/vall-e

PyTorch implementation of VALL-E(Zero-Shot Text-To-Speech), Reproduced Demo https://lifeiteng.github.io/valle/index.html

Repository: https://github.com/lifeiteng/vall-e
Canonical: https://ross.abutalabs.com/products/lifeiteng-vall-e
Homepage: https://lifeiteng.github.io/valle/index.html
Language: Python
License: Apache-2.0
License Family: permissive
Topics: in-context-learning, large-language-models, text-to-speech, tts, chatgpt, vall-e, valle
Last push: 2025-09-10T16:08:02+00:00

## Health v2 (maintenance only)
Score: 40/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 41, release rhythm 8, longevity 93
- inputs: {"age_days": 1314, "days_push": 357, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2215, forks 329 (observed 2026-08-28T04:06:27.079559+00:00)

## What it is
An unofficial PyTorch implementation of VALL-E, a zero-shot text-to-speech model that treats TTS as a conditional language modeling task over neural codec codes. It supports training on a single GPU and reproduces the paper's results with demo audio for English and Chinese.

## Use cases
- synthesize speech in a cloned voice from a 3-second audio prompt
- train a zero-shot text-to-speech model on my own dataset
- reproduce the VALL-E paper results
- convert text to speech with emotion and acoustic environment preserved
- experiment with neural codec language models for audio generation

## When to choose
- you want an open-source, trainable VALL-E implementation in PyTorch
- you need zero-shot TTS with voice cloning from short prompts
- you want to train TTS on a single consumer GPU

## When to avoid
- you need a production-ready pretrained model or hosted service (none is provided to prevent misuse)
- you need a simple pip-install TTS API without training setup
- you require official support or guaranteed parity with the original model

## Facets
- artifact type: library
- maturity: active
- function: tts, machine-learning, deep-learning, llm-training
- domain: speech-processing, deep-learning, machine-learning
- platform: python
- tags: pytorch, zero-shot-tts, neural-codec, voice-cloning, speech-synthesis, vall-e, in-context-learning, natural-language-processing, linux, macos, gpu

## Member repositories
- lifeiteng/vall-e (main) score 40

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:27.079559+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 2026-08-30T02:45:58.336235+00:00, confidence not recorded.
  - readme: https://github.com/lifeiteng/vall-e (fetched 2026-08-28T04:06:27.079559+00:00, sha 09784d3d0b58)
  - homepage: https://lifeiteng.github.io/valle/index.html (fetched 2026-08-29T10:26:26.866290+00:00, sha ab5236f4cbcd)
- Data as of 2026-08-30T08:39:29.467469+00:00.