This model was trained by ftshijt using vctk recipe in espnet. Follow the ESPnet installation instructions if you haven't done that already.
SAVRN Model Hub · Models by Task
Text to Speech Models
64 open-weight text to speech models in the SAVRN Model Hub, with Qwen, OpenMOSS and AI at Meta publishing the most.
64Models
42Publishers
10M to 8.5BParameter range
7Licenses
SAVRN's Take
Nobody buys a rack for text to speech. The 31 models on this page with a parameter count run from a 10 million parameter voice tuner for Kokoro to OpenMOSS's 8.5 billion parameter MOSS-TTS-v1.5, which wants 20.4 GB at 16-bit and 5.1 GB at 4-bit. Qwen's 1.9 billion parameter Qwen3-TTS-12Hz-1.7B-CustomVoice fits in 4.6 GB. Every indexed entry prices out at one MI300X at $1.85 an hour, so the GPU is not where the cost lives. What differs is the job: XTTS-v2 clones a voice from a 6-second clip across 17 languages, Qwen3-TTS takes tone and speaking rate instructions across 10, OmniVoice lists more than 600, and MOSS-TTS-v1.5 pairs zero-shot cloning with long-form generation on a 40,960 token context.
Downloads favor the small end. Hexgrad's Kokoro-82M, 82 million parameters under Apache 2.0, pulled 11,687,399 in a month, ahead of Coqui's XTTS-v2 at 7,324,519 and audio.cpp-gguf at 4,129,715, GGUF conversions for that runtime rather than a model. Qwen, OpenMOSS and AI at Meta hold four entries each.
Licensing is where we would spend the diligence. Apache 2.0 covers 32 of the 64 and MIT another 10. Nine carry CC BY-NC 4.0, F5-TTS and its 984,345 downloads among them, and the NC means noncommercial. Seven are filed as other, XTTS-v2 and audio.cpp-gguf included, two state no license, and OmniVoice's field is blank. Read the license text on the publisher's repository for anything outside Apache and MIT, and pick the runtime before the model: coqui, chatterbox, f5-tts, omnivoice and audio.cpp are each their own library, and that choice outlasts any checkpoint.
Most Downloaded
| Model | Publisher | Parameters | License | Monthly downloads | Cheapest GPUs at 16-bit |
|---|---|---|---|---|---|
| Kokoro-82M | Hexgrad | — | apache-2.0 | 11.7M | — |
| XTTS-v2 | Coqui.ai | — | other | 7.3M | — |
| audio.cpp-gguf | Audio.cpp | — | other | 4.1M | — |
| Qwen3-TTS-12Hz-1.7B-CustomVoice | Qwen | 1.9B | apache-2.0 | 2.6M | 1x MI300X, $1.85/hr |
| chatterbox | Resemble AI | — | mit | 1.9M | — |
| OmniVoice | K2 FSA | 613M | Not stated | 1.3M | 1x MI300X, $1.85/hr |
| Qwen3-TTS-12Hz-0.6B-CustomVoice | Qwen | 906M | apache-2.0 | 995.5k | 1x MI300X, $1.85/hr |
| F5-TTS | Yushen CHEN | — | cc-by-nc-4.0 | 984.3k | — |
| Kokoro-82M-v1.0-ONNX | ONNX Community | — | apache-2.0 | 739.9k | — |
| Qwen3-TTS-12Hz-0.6B-Base | Qwen | 915M | apache-2.0 | 641.3k | 1x MI300X, $1.85/hr |
Licenses
| License | Models | Commercial use |
|---|---|---|
| apache-2.0 | 32 | Yes |
| mit | 10 | Yes |
| cc-by-nc-4.0 | 9 | Not without separate permission |
| other | 7 | Read the license |
| cc-by-4.0 | 3 | Yes |
| not stated | 2 | Not stated |
| gpl-3.0 | 1 | Read the license |
Who Publishes Them
| Publisher | Models |
|---|---|
| Qwen | 4 |
| OpenMOSS | 4 |
| AI at Meta | 4 |
| Pham Nguyen Ngoc Bao | 3 |
| MyShell.ai | 3 |
| Microsoft | 3 |
All 64 Models, Page 2 of 2
English grapheme-to-phoneme resources for KittenTTS and Kokoro TTS. Downloaded automatically by mlx-audio-swift when using KittenTTS or Kokoro TTS with English voices. Ported from Misaki G2P engine (MIT license).
This model was converted to MLX format from hexagrad/Kokoro-82M using mlx-audio version 0.0.1. Refer to the original model card for more details on the model.
中文语音合成(TTS)模型,支持语音克隆与数字/中英混读,CPU 优先设计,权重原生 int8 量化。 基于 Kyutai Pocket TTS 架构:6 层 FlowLM(dmodel 1024)+ Mimi 神经音频编解码器(24 kHz),端到端流式推理。本版本(R21)相对上一代的核心变化: - 原生 int8 交付:FlowLM 注意力与 FFN 权重已离线量化打包,无需运行时转换,体积 223 MB,CPU 速度提升约 1 倍 - 9174 token 词表:官方英文 BPE-4000 基底 + 追加单字汉字/全角标点,bytefallback 兜底生僻字符 - 中英混读:APP、VIP、5G、WiFi 等英文术语按中文习惯读法 - 模型为预量化 int8 权重(modelint8.pt),加载即用,无需 quantize=True 或有界面演示(需 pip install gradio): - 采样温度 temp=0.3(默认);文本越长对参考音频质量越敏感 文本准确性用 whisper-small 对合成音频转写后计算字错误率(CER),每个场景 10 次独立采样取均值,生产配置(单步采样)。数值越低越好。 测试环境:Intel i7-12700KF,生成 52 字客服长句(约 8.5 秒音频),每档 3 次取均值。 参考音频:客服目标音色(kefulong / digitsorder / codeswitch / shortunseen)与 AISHELL-3 说话人 A/B(speakerAlong /…
Questions
Which Text to speech models are most downloaded?
By monthly downloads reported by the Hugging Face Hub: vctk_tts_train_espnet_rawnet_vits (1).