# BytedanceSpeech/seed-tts-eval

Repository: https://github.com/BytedanceSpeech/seed-tts-eval
Canonical: https://ross.abutalabs.com/products/seed-tts-eval
Language: Python
License Family: other
Last push: 2024-06-14T09:47:14+00:00

## Health v2 (maintenance only)
Score: 24/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 58
- inputs: {"age_days": 821, "days_push": 810, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1595, forks 147 (observed 2026-08-28T04:05:09.118611+00:00)

## What it is
An objective evaluation test set and metric scripts from ByteDance's seed-TTS project for benchmarking zero-shot text-to-speech and voice conversion. It includes English and Mandarin test samples plus code for computing word error rate (WER) and speaker similarity (SIM).

## Use cases
- evaluate zero-shot TTS models on English and Mandarin
- compute WER for synthesized speech with Whisper and Paraformer
- measure speaker similarity of generated audio with WavLM embeddings
- benchmark voice conversion quality
- compare TTS systems on hard-case Mandarin samples

## When to choose
- you need a standard objective benchmark for zero-shot TTS or VC models
- you want reproducible WER and SIM evaluation scripts
- you are evaluating multilingual (EN/ZH) speech generation

## When to avoid
- you want the seed-TTS model itself, which is not released
- you need perceptual or MOS-style subjective evaluation
- you need languages other than English and Mandarin

## Facets
- artifact type: dataset
- maturity: stable
- function: speech-recognition, benchmarking, audio-processing
- domain: speech-processing, machine-learning, artificial-intelligence
- platform: python, cli
- tags: tts-evaluation, zero-shot-tts, voice-conversion, wer, speaker-similarity, test-set

## Member repositories
- BytedanceSpeech/seed-tts-eval (main) score 24

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:09.118611+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 2026-08-30T03:54:04.895964+00:00, confidence not recorded.
  - readme: https://github.com/BytedanceSpeech/seed-tts-eval (fetched 2026-08-28T04:05:09.118611+00:00, sha 358d7f01d205)
- Data as of 2026-08-30T08:39:29.467469+00:00.