ONNX 模型库
返回模型

说明文档


license: apache-2.0 library_name: transformers.js language:

  • en base_model:
  • hexgrad/Kokoro-82M pipeline_tag: text-to-speech

Kokoro TTS

Kokoro 是一个拥有 8200 万参数的前沿 TTS(文本转语音)模型(文本输入/音频输出)。

目录

使用方法

JavaScript

首先,从 NPM 安装 kokoro-js 库:

npm i kokoro-js

然后你可以按以下方式生成语音:

import { KokoroTTS } from "kokoro-js";

const model_id = "onnx-community/Kokoro-82M-ONNX";
const tts = await KokoroTTS.from_pretrained(model_id, {
  dtype: "q8", // 选项: "fp32", "fp16", "q8", "q4", "q4f16"
});

const text = "Life is like a box of chocolates. You never know what you're gonna get.";
const audio = await tts.generate(text, {
  // 使用 `tts.list_voices()` 列出所有可用的声音
  voice: "af_bella",
});
audio.save("audio.wav");

Python

import os
import numpy as np
from onnxruntime import InferenceSession

# 你可以按以下方式生成 token id:
#   1. 使用 https://github.com/hexgrad/misaki 将输入文本转换为音素
#   2. 使用 https://huggingface.co/hexgrad/Kokoro-82M/blob/785407d1adfa7ae8fbef8ffd85f34ca127da3039/config.json#L34-L148 将音素映射为 id
tokens = [50, 157, 43, 135, 16, 53, 135, 46, 16, 43, 102, 16, 56, 156, 57, 135, 6, 16, 102, 62, 61, 16, 70, 56, 16, 138, 56, 156, 72, 56, 61, 85, 123, 83, 44, 83, 54, 16, 53, 65, 156, 86, 61, 62, 131, 83, 56, 4, 16, 54, 156, 43, 102, 53, 16, 156, 72, 61, 53, 102, 112, 16, 70, 56, 16, 138, 56, 44, 156, 76, 158, 123, 56, 16, 62, 131, 156, 43, 102, 54, 46, 16, 102, 48, 16, 81, 47, 102, 54, 16, 54, 156, 51, 158, 46, 16, 70, 16, 92, 156, 135, 46, 16, 54, 156, 43, 102, 48, 4, 16, 81, 47, 102, 16, 50, 156, 72, 64, 83, 56, 62, 16, 156, 51, 158, 64, 83, 56, 16, 44, 157, 102, 56, 16, 44, 156, 76, 158, 123, 56, 4]

# 上下文长度为 512,但需要在开头和结尾留出 pad token 0 的位置
assert len(tokens) <= 510, len(tokens)

# 基于 len(tokens) 的风格向量,ref_s 的形状为 (1, 256)
voices = np.fromfile('./voices/af.bin', dtype=np.float32).reshape(-1, 1, 256)
ref_s = voices[len(tokens)]

# 添加 pad id,并重塑 tokens,现在形状应为 (1, <=512)
tokens = [[0, *tokens, 0]]

model_name = 'model.onnx' # 选项: model.onnx, model_fp16.onnx, model_quantized.onnx, model_q8f16.onnx, model_uint8.onnx, model_uint8f16.onnx, model_q4.onnx, model_q4f16.onnx
sess = InferenceSession(os.path.join('onnx', model_name))

audio = sess.run(None, dict(
    input_ids=tokens,
    style=ref_s,
    speed=np.ones(1, dtype=np.float32),
))[0]

可选,将音频保存到文件:

import scipy.io.wavfile as wavfile
wavfile.write('audio.wav', 24000, audio[0])

声音/示例

Life is like a box of chocolates. You never know what you're gonna get.

名称 国籍 性别 示例
af_heart 美国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/S_9tkA75BT_QHKOzSX6S-.wav"></audio>
af_alloy 美国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/wiZ3gvlL--p5pRItO4YRE.wav"></audio>
af_aoede 美国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/Nv1xMwzjTdF9MR8v0oEEJ.wav"></audio>
af_bella 美国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/sWN0rnKU6TlLsVdGqRktF.wav"></audio>
af_jessica 美国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/2Oa4wITWAmiCXJ_Q97-7R.wav"></audio>
af_kore 美国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/AOIgyspzZWDGpn7oQgwtu.wav"></audio>
af_nicole 美国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/EY_V2OGr-hzmtTGrTCTyf.wav"></audio>
af_nova 美国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/X-xdEkx3GPlQG5DK8Gsqd.wav"></audio>
af_river 美国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/ZqaV2-xGUZdBQmZAF1Xqy.wav"></audio>
af_sarah 美国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/xzoJBl1HCvkE8Fl8Xu2R4.wav"></audio>
af_sky 美国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/ubebYQoaseyQk-jDLeWX7.wav"></audio>
am_adam 美国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/tvauhDVRGvGK98I-4wv3H.wav"></audio>
am_echo 美国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/qy_KuUB0hXsu-u8XaJJ_Z.wav"></audio>
am_eric 美国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/JhqPjbpMhraUv5nTSPpwD.wav"></audio>
am_fenrir 美国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/c0R9caBdBiNjGUUalI_DQ.wav"></audio>
am_liam 美国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/DFHvulaLeOjXIDKecvNG3.wav"></audio>
am_michael 美国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/IPKhsnjq1tPh3JmHH8nEg.wav"></audio>
am_onyx 美国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/ov0pFDfE8NNKZ80LqW6Di.wav"></audio>
am_puck 美国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/MOC654sLMHWI64g8HWesV.wav"></audio>
am_santa 美国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/LzA6JmHBvQlhOviy8qVfJ.wav"></audio>
bf_alice 英国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/9mnYZ3JWq7f6U12plXilA.wav"></audio>
bf_emma 英国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/_fvGtKMttRI0cZVGqxMh8.wav"></audio>
bf_isabella 英国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/VzlcJpqGEND_Q3duYnhiu.wav"></audio>
bf_lily 英国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/qZCoartohiRlVamY8Xpok.wav"></audio>
bm_daniel 英国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/Eb0TLnLXHDRYOA3TJQKq3.wav"></audio>
bm_fable 英国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/NT9XkmvlezQ0FJ6Th5hoZ.wav"></audio>
bm_george 英国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/y6VJbCESszLZGupPoqNkF.wav"></audio>
bm_lewis 英国 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/RlB5BRvLt-IFvTjzQNxCh.wav"></audio>

量化版本

该模型对量化具有很好的适应性,能够以原始模型大小的一小部分实现高效的高质量语音合成。

How could I know? It's an unanswerable question. Like asking an unborn child if they'll lead a good life. They haven't even been born.

模型 大小 (MB) 示例
model.onnx (fp32) 326 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/njexBuqPzfYUvWgs9eQ-_.wav"></audio>
model_fp16.onnx (fp16) 163 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/8Ebl44hMQonZs4MlykExt.wav"></audio>
model_quantized.onnx (8-bit) 92.4 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/9SLOt6ETclZ4yRdlJ0VIj.wav"></audio>
model_q8f16.onnx (Mixed precision) 86 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/gNDMqb33YEmYMbAIv_Grx.wav"></audio>
model_uint8.onnx (8-bit & mixed precision) 177 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/tpOWRHIWwEb0PJX46dCWQ.wav"></audio>
model_uint8f16.onnx (Mixed precision) 114 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/vtZhABzjP0pvGD7dRb5Vr.wav"></audio>
model_q4.onnx (4-bit matmul) 305 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/8FVn0IJIUfccEBWq8Fnw_.wav"></audio>
model_q4f16.onnx (4-bit matmul & fp16 weights) 154 <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/7DrgWC_1q00s-wUJuG44X.wav"></audio>

chae12121/kokorotest

作者 chae12121

text-to-speech transformers.js
↓ 1 ♥ 0

创建时间: 2026-01-09 15:19:43+00:00

更新时间: 2026-01-09 15:35:07+00:00

在 Hugging Face 上查看

文件 (69)

.gitattributes
README.md
config.json
onnx/model.onnx ONNX
onnx/model_fp16.onnx ONNX
onnx/model_q4.onnx ONNX
onnx/model_q4f16.onnx ONNX
onnx/model_q8f16.onnx ONNX
onnx/model_quantized.onnx ONNX
onnx/model_uint8.onnx ONNX
onnx/model_uint8f16.onnx ONNX
tokenizer.json
tokenizer_config.json
voices.json
voices/af.bin
voices/af_alloy.bin
voices/af_aoede.bin
voices/af_bella.bin
voices/af_heart.bin
voices/af_jessica.bin
voices/af_kore.bin
voices/af_nicole.bin
voices/af_nova.bin
voices/af_river.bin
voices/af_sarah.bin
voices/af_sky.bin
voices/am_adam.bin
voices/am_echo.bin
voices/am_eric.bin
voices/am_fenrir.bin
voices/am_liam.bin
voices/am_michael.bin
voices/am_onyx.bin
voices/am_puck.bin
voices/am_santa.bin
voices/bf_alice.bin
voices/bf_emma.bin
voices/bf_isabella.bin
voices/bf_lily.bin
voices/bm_daniel.bin
voices/bm_fable.bin
voices/bm_george.bin
voices/bm_lewis.bin
voices/ef_dora.bin
voices/em_alex.bin
voices/em_santa.bin
voices/ff_siwis.bin
voices/hf_alpha.bin
voices/hf_beta.bin
voices/hm_omega.bin
voices/hm_psi.bin
voices/if_sara.bin
voices/im_nicola.bin
voices/jf_alpha.bin
voices/jf_gongitsune.bin
voices/jf_nezumi.bin
voices/jf_tebukuro.bin
voices/jm_kumo.bin
voices/pf_dora.bin
voices/pm_alex.bin
voices/pm_santa.bin
voices/zf_xiaobei.bin
voices/zf_xiaoni.bin
voices/zf_xiaoxiao.bin
voices/zf_xiaoyi.bin
voices/zm_yunjian.bin
voices/zm_yunxi.bin
voices/zm_yunxia.bin
voices/zm_yunyang.bin