Skip to content

最强的声音克隆

Python · active 2026-01-222026-08-10 (UTC)

Partial coverage11,201 / 11,856 hourly files (94%) · 2 absent upstream · 653 failed, retryable2025-04-032026-08-10 (UTC)— sampled evenly across the window, so rankings and trends hold; absolute counts scale up.
Events
4K
Pushes
5
Pull requests
14
Issues
48
Stars
3.3K
Forks
415

Activity over time

Daily event counts in the loaded window

Line chart, 201 days from 2026-01-22 to 2026-08-10. Pushes: 5 total, peak 2 in a day. Pull requests: 14 total, peak 2 in a day. Issues: 48 total, peak 9 in a day. Comments: 135 total, peak 11 in a day. Stars: 3,320 total, peak 571 in a day.

  • Pushes
  • Pull requests
  • Issues
  • Comments
  • Stars

Stars, PRs, issues and forks are under-captured in the later part of this window. GH Archive progressively stopped capturing non-push events during 2026 — −95% or worse by the end of the window. Every series here except Pushes fades for that reason, so a decline above reflects the archive, not this repository. Pushes stay reliable throughout, so read them, and the contributor counts derived from them, as the real signal. Data health has the measurements.

Top contributors

Pushes, PRs, issues, reviews and comments — stars and forks excluded, so this is contribution rather than popularity

Recent activity

Latest issues, pull requests and releases

  • Issue#39github-actions[bot]2026-06-27 09:03
    Fine-tuned model outputs noise despite decreasing training loss
  • Issue comment#39github-actions[bot]2026-06-27 09:03
    Fine-tuned model outputs noise despite decreasing training loss
  • Issue comment#323arpit7632026-06-10 05:34
    Fine-tuning Qwen3-TTS-12Hz-1.7B-Base for Bengali: Foreign/Hindi Accent Bias and Lack of Naturalness
  • Issue#305github-actions[bot]2026-06-07 09:12
    BASE模型,参考音频为英文读数字会有问题
  • Issue#179github-actions[bot]2026-05-25 09:32
    Finetuning Base results in progressively faster speech with every epoch
  • Issue comment#215github-actions[bot]2026-05-17 08:51
    Refactor/add training functions
  • Issue comment#263github-actions[bot]2026-05-17 08:51
    Unstable inference when using vllm with cloning
  • Issue#277github-actions[bot]2026-05-09 08:41
    数字读错
  • Issue#89github-actions[bot]2026-05-05 08:55
    Very slow inference on 5090
  • Issue#256github-actions[bot]2026-05-04 09:02
    克隆语音生成 996 读错
  • Issue comment#10github-actions[bot]2026-05-03 08:42
    Streaming audio
  • Issue comment#34NielsRogge2026-05-02 10:44
    Release Qwen3-TTS 25Hz models and tokenizer on Hugging Face
  • Issue comment#248github-actions[bot]2026-05-02 08:39
    Instruction following works for emotions but fails for specific Chinese dialects (Sichuan/Shanghai/Cantonese)
  • Issue#118github-actions[bot]2026-05-02 08:39
    EOS token not being found on pretrained base models
  • Issue comment#118github-actions[bot]2026-05-02 08:39
    EOS token not being found on pretrained base models
  • Issue comment#260github-actions[bot]2026-04-28 09:02
    TypeError: cannot pickle 'dict_keys' object when loading 4-bit quantized model (divyajot5005/Qwen3-TTS-12Hz-1.7B-Base-BNB-4bit)
  • Issue comment#265github-actions[bot]2026-04-27 09:00
    qwen3-tts-instruct-flash can not pronounce numbers in English, but qwen3-tts-flash works, that's weird
  • Issue#194github-actions[bot]2026-04-26 08:34
    The training methodology of the base model is inconsistent with the description provided in the paper.
  • Issue#39github-actions[bot]2026-04-22 08:45
    Fine-tuned model outputs noise despite decreasing training loss
  • Issue comment#145bruce22332026-04-21 16:49
    Update Python Dependency
  • Issue#145github-actions[bot]2026-04-19 08:28
    Update Python Dependency
  • Issue#68github-actions[bot]2026-04-18 08:21
    测试发现的几个问题,避免大家踩坑,距离生产可用还有一段距离
  • Issue comment#141billsikalong2026-04-16 02:14
    How to generate Cantonese (粤语)?
  • Issue comment#240github-actions[bot]2026-04-15 08:43
    建议把 dialect language 的限制去掉,以便于用户可以在别的音色下使用方言语言
  • Pull request#295ElhamDevelopmentStudio2026-04-14 10:09

Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 3,320 stars here means stars gained during the window, not the repo's star count.