简介

Kokoro 82M 是一款极其轻量级且高效的文本转语音（TTS）模型。虽然参数量仅为 82M，但其音质表现出人意料地自然，远超传统的轻量级合成方案。该版本采用 ONNX 格式，这意味着它脱离了沉重的深度学习框架，可以在 CPU 上实现极速推理，非常适合开发者将其集成到本地应用、边缘设备或低功耗服务器中。对于追求低延迟、无需 GPU 资源且对语音自然度有一定要求的中文/英文场景，它是目前一个极具性价比的替代方案。

核心亮点

极小参数量实现高自然度语音合成
ONNX 格式支持，CPU 推理速度极快
无需高端 GPU，部署门槛极低
适合集成至本地软件或轻量级 AI 助手

使用方法

安装依赖

# 安装 Hugging Face transformers
pip install transformers torch

SDK 使用

# 使用 transformers 加载模型
from transformers import AutoModel, AutoTokenizer

model = AutoModel.from_pretrained("onnx-community/Kokoro-82M-v1.0-ONNX")
tokenizer = AutoTokenizer.from_pretrained("onnx-community/Kokoro-82M-v1.0-ONNX")

Hugging Face 下载

我们推荐使用命令行或者 Hugging Face Hub SDK 来进行模型的下载。

操作指引：在下载前，请先通过如下命令安装 huggingface_hub：

操作指引

pip install -U huggingface_hub

命令行下载

下载完整模型库

huggingface-cli download onnx-community/Kokoro-82M-v1.0-ONNX

下载单个文件到指定本地文件夹（以下载 config.json 到当前路径下 ./dir 目录为例）

huggingface-cli download onnx-community/Kokoro-82M-v1.0-ONNX config.json --local-dir ./dir

更多命令行下载选项，可参见官方文档

SDK 下载

# 模型下载
from huggingface_hub import snapshot_download
model_dir = snapshot_download('onnx-community/Kokoro-82M-v1.0-ONNX')

Git 下载

请确保 lfs 已经被正确安装

Git 下载

git lfs install
git clone https://huggingface.co/onnx-community/Kokoro-82M-v1.0-ONNX

如果您希望跳过 lfs 大文件下载，可以使用如下命令

跳过 LFS

GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/onnx-community/Kokoro-82M-v1.0-ONNX

模型文件托管在 Hugging Face Hub，使用 HF CLI / SDK / Git 直接下载，不经过本站。

PyTorch / Transformers 使用

安装 Transformers

pip install -U transformers torch

模型加载和推理

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained('onnx-community/Kokoro-82M-v1.0-ONNX')
tokenizer = AutoTokenizer.from_pretrained('onnx-community/Kokoro-82M-v1.0-ONNX')

模型下载

我们推荐使用命令行或者 ModelScope SDK 来进行模型的下载。

操作指引：在下载前，请先通过如下命令安装 ModelScope：

操作指引

pip install modelscope

命令行下载

下载完整模型库

modelscope download --model onnx-community/Kokoro-82M-v1.0-ONNX

下载单个文件到指定本地文件夹（以下载 README.md 到当前路径下 dir 目录为例）

modelscope download --model onnx-community/Kokoro-82M-v1.0-ONNX README.md --local_dir ./dir

更多更丰富的命令行下载选项，可参见具体文档

SDK 下载

# 模型下载
from modelscope import snapshot_download
model_dir = snapshot_download('onnx-community/Kokoro-82M-v1.0-ONNX')

Git 下载

请确保 lfs 已经被正确安装

Git 下载

git lfs install
git clone https://www.modelscope.cn/onnx-community/Kokoro-82M-v1.0-ONNX.git

如果您希望跳过 lfs 大文件下载，可以使用如下命令

跳过 LFS

GIT_LFS_SKIP_SMUDGE=1 git clone https://www.modelscope.cn/onnx-community/Kokoro-82M-v1.0-ONNX.git

ModelScope 模型页直接下载模型文件；无需将模型文件放在本站服务器。

Notebook 快速开发

下载并安装 ModelScope library

pip install "modelscope[audio,cv,nlp,multi-modal,science]" -f https://modelscope.oss-cn-beijing.aliyuncs.com/releases/repo.html

模型加载和推理

from modelscope.pipelines import pipeline
from modelscope.utils.constant import Tasks

p = pipeline('text-generation', 'onnx-community/Kokoro-82M-v1.0-ONNX')

完整文档

来源: HuggingFace

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

en

base_model:

hexgrad/Kokoro-82M

pipeline_tag: text-to-speech
---

Kokoro TTS

Kokoro is a frontier TTS model for its size of 82 million parameters (text in/audio out).

Usage

JavaScript

First, install the kokoro-js library from NPM using:

bash

npm i kokoro-js

You can then generate speech as follows:

js

import { KokoroTTS } from "kokoro-js";
const model_id = "onnx-community/Kokoro-82M-ONNX";
const tts = await KokoroTTS.from_pretrained(model_id, {
  dtype: "q8", // Options: "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, {
  // Use tts.list_voices() to list all available voices
  voice: "af_bella",
});
audio.save("audio.wav");

Python

python

import os
import numpy as np
from onnxruntime import InferenceSession
You can generate token ids as follows:
  1. Convert input text to phonemes using https://github.com/hexgrad/misaki
  2. Map phonemes to ids using https://huggingface.co/hexgrad/Kokoro-82M/blob/785407d1adfa7ae8fbef8ffd85f34ca127da3039/config.json#L34-L148
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]
Context length is 512, but leave room for the pad token 0 at the start & end
assert len(tokens) <= 510, len(tokens)
Style vector based on len(tokens), ref_s has shape (1, 256)
voices = np.fromfile('./voices/af.bin', dtype=np.float32).reshape(-1, 1, 256)
ref_s = voices[len(tokens)]
Add the pad ids, and reshape tokens, should now have shape (1, <=512)
tokens = [[0, *tokens, 0]]
model_name = 'model.onnx' # Options: 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]

Optionally, save the audio to a file:

py

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

Voices/Samples

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

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

Kokoro 82M v1.0 ONNX

简介

核心亮点

使用方法

Hugging Face 下载

命令行下载

SDK 下载

Git 下载

PyTorch / Transformers 使用

模型下载

命令行下载

SDK 下载

Git 下载

Notebook 快速开发

完整文档

Kokoro TTS

Table of contents

Usage

JavaScript

Python

You can generate token ids as follows:

1. Convert input text to phonemes using https://github.com/hexgrad/misaki

2. Map phonemes to ids using https://huggingface.co/hexgrad/Kokoro-82M/blob/785407d1adfa7ae8fbef8ffd85f34ca127da3039/config.json#L34-L148

Context length is 512, but leave room for the pad token 0 at the start & end

Style vector based on len(tokens), ref_s has shape (1, 256)

Add the pad ids, and reshape tokens, should now have shape (1, <=512)

Voices/Samples

扫码进交流群