> ## Content Index
> Fetch the complete content index at: https://www.hubwiz.com/blog/llms.txt
> Use this file to discover other available public pages before exploring further.

# VoxCPM2 多语种开源TTS模型
- URL: https://www.hubwiz.com/blog/voxcpm2-tts/
- Published: 2026-05-09T23:45:33.000Z
- Updated: 2026-05-09T23:45:33.000Z
- Description: 30种语言，48kHz输出，上下文感知的韵律。从文本设计任何声音。从短音频片段克隆并带有风格引导。无需语言标签。一切都在本地运行。免费用于商业用途。
- Author: admin
- Tags: MODEL-ZOO, VOXCPM2, #Import 2026-10-14 14:41

我已经介绍过几款优秀的TTS模型，包括[Qwen3-TTS](https://medium.com/@bytefer/the-free-open-source-alternative-to-elevenlabs-is-finally-here-3b97edd63e2a?ref=hubwiz.com)、Fish Audio S2 Pro和Voxtral TTS。最近，OpenBMB团队发布了最新的**VoxCPM2**模型，拥有20亿参数，支持语音克隆、语音设计和高质量语音合成，支持30种主流语言，包括英语、中文、日语、韩语、德语和法语。

VoxCPM2特性：

- 支持30种语言，无需语言标签，直接输入任何支持语言的文本。
- 支持语音设计，输入自然语言描述即可生成新声音，无需参考音频。
- 支持语音克隆，带有风格引导来控制情感、语速和表现力。
- 支持生成高质量48kHz音频。

## 1、VoxCPM架构

VoxCPM采用无分词器的扩散自回归架构，在连续潜空间而非离散标记中对语音进行建模。

![None](https://www.hubwiz.com/blog/content/images/2026/05/0-bD_0srlzCFW8ni5w.png)

[https://voxcpm.readthedocs.io/](https://voxcpm.readthedocs.io/?ref=hubwiz.com)

## 2、本地部署

VoxCPM2官方文档已经详细介绍了如何使用PyTorch和CUDA运行VoxCPM2模型。接下来，我将介绍如何使用**mlx-audio**在macOS上本地部署VoxCPM2模型。

配置虚拟环境

```
uv venv .venv
source .venv/bin/activate

```

安装mlx-audio和soundfile

```
uv pip install "git+https://github.com/Blaizzy/mlx-audio" --prerelease=allow
uv pip install soundfile

```

下载模型。

你可以根据你的电脑配置和实际需求下载对应的量化模型。

![None](https://www.hubwiz.com/blog/content/images/2026/05/1-56qUjzvXZyyH7MHu0BExTA.png)

mlx voxcpm2 模型

```
hf download mlx-community/VoxCPM2-4bit --local-dir ./models/VoxCPM2-4bit
# or
hf download mlx-community/VoxCPM2-8bit --local-dir ./models/VoxCPM2-8bit
# or
hf download mlx-community/VoxCPM2-bf16 --local-dir ./models/VoxCPM2-bf16

```

![None](https://www.hubwiz.com/blog/content/images/2026/05/1-TVZ4UdIipb0ME4z9tDCPUQ.png)

mlx voxcpm2 模型

## 3、零样本生成

```
import numpy as np
from mlx_audio.tts.utils import load
from mlx_audio.audio_io import write as audio_write

MODEL_DIR = "models/VoxCPM2-8bit"
OUTPUT_PATH = "zero_shot.wav"
model = load(MODEL_DIR)

result = next(model.generate("Hello, this is VoxCPM2 on Apple Silicon."))
audio_mx = result.audio

audio_write(
    str(OUTPUT_PATH),
    np.array(audio_mx),
    model.sample_rate,
    format="wav",
)

```

需要注意的是，如果要合成的文本包含括号，需要进行转义；否则无法正确合成语音。

## 4、语音设计

```
import numpy as np
from mlx_audio.tts.utils import load
from mlx_audio.audio_io import write as audio_write

MODEL_DIR = "models/VoxCPM2-8bit"
OUTPUT_PATH = "voice_design.wav"
model = load(MODEL_DIR)

result = next(model.generate(
    text="Hello, welcome to VoxCPM2.",
    instruct="A young woman, warm and gentle voice",
))
audio_mx = result.audio

audio_write(
    str(OUTPUT_PATH),
    np.array(audio_mx),
    model.sample_rate,
    format="wav",
)

```

## 5、语音克隆

```
import numpy as np
from mlx_audio.tts.utils import load
from mlx_audio.audio_io import write as audio_write

MODEL_DIR = "models/VoxCPM2-8bit"
OUTPUT_PATH = "voice_cloning.wav"
model = load(MODEL_DIR)

result = next(model.generate(
    text="Hello, this is VoxCPM2 on Apple Silicon.",
    ref_audio="lisa.wav",
))
audio_mx = result.audio

audio_write(
    str(OUTPUT_PATH),
    np.array(audio_mx),
    model.sample_rate,
    format="wav",
)

```

## 6、终极克隆

对于涉及有声书等长内容制作的场景，你需要同时提供参考音频和对应的转录文本以确保一致性。

```
import numpy as np
from mlx_audio.tts.utils import load
from mlx_audio.audio_io import write as audio_write

MODEL_DIR = "models/VoxCPM2-8bit"
OUTPUT_PATH = "ultimate_cloning.wav"
model = load(MODEL_DIR)

result = next(model.generate(
    text="2B-parameter multilingual tokenizer-free TTS model with 48kHz studio-quality output. Supports zero-shot generation, voice design, voice cloning, and continuation for long-form speech. 30 languages including English, Chinese, Indonesian, Japanese, Korean, and more.",
    prompt_text="VoxCPM2 is a tokenizer-free, diffusion autoregressive Text-to-Speech model",
    prompt_audio="lisa.wav",
))

audio_mx = result.audio

audio_write(
    str(OUTPUT_PATH),
    np.array(audio_mx),
    model.sample_rate,
    format="wav",
)

```

## 7、结束语

VoxCPM2是一个强大的TTS模型。如果你有语音合成需求，可以评估其能力。如果它不能满足你的需求，你可以测试Qwen3-TTS，它提供0.6B和1.7B两种规模，支持10种主流语言，包括英语、中文、日语、韩语、德语和法语。它也支持语音合成、语音设计和语音克隆，但你需要在不同的模型之间切换。

---

原文链接: [The Free, Open-Source Alternative to ElevenLabs Is Finally Here](https://medium.com/@bytefer/the-free-open-source-alternative-to-elevenlabs-is-finally-here-3cbbacd9c6b9?ref=hubwiz.com)

汇智网翻译整理，转载请标明出处