VoxCPM2-MLX / README.md
jennyxue's picture
Model card and the original license
6fca4c3 verified
|
Raw History Blame Contribute Delete
18.7 kB
---
base_model: openbmb/VoxCPM2
base_model_relation: quantized
license: apache-2.0
pipeline_tag: text-to-speech
library_name: mlx-audio
tags:
- mlx
- moonbox
- mlx-audio
- voxcpm
- voxcpm2
- text-to-speech
- voice-cloning
- speech
---
# VoxCPM2 MLX
MLX builds of [VoxCPM2](https://huggingface.co/openbmb/VoxCPM2), made by Moonbox for Apple Silicon Macs. Each precision is a folder; keep a folder's
files together. Only the parts listed below are converted: the rest of the pipeline comes from the original
repository unless the section below says otherwise.
## Files
The last column is how far each build is from the original PyTorch model on the same inputs (see the section below
for what was measured).
| Precision | File | Size | sha256 | Checked against the original |
|---|---|---|---|---|
| bf16 | [`bf16/config.json`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/bf16/config.json) | 6.6 kB | `0b5bee352207` | WER 0.0 % / CER 0.0 %, speaker sim 0.973 / 0.967, DNSMOS 3.12 / 2.33 (original 0.0 % / 0.0 %, 0.972 / 0.964, 2.93 / 2.28) |
| bf16 | [`bf16/model.safetensors`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/bf16/model.safetensors) | 4.77 GB | `8ec57a6a9761` | WER 0.0 % / CER 0.0 %, speaker sim 0.973 / 0.967, DNSMOS 3.12 / 2.33 (original 0.0 % / 0.0 %, 0.972 / 0.964, 2.93 / 2.28) |
| bf16 | [`bf16/model.safetensors.index.json`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/bf16/model.safetensors.index.json) | 65.7 kB | `1a4680fb582f` | WER 0.0 % / CER 0.0 %, speaker sim 0.973 / 0.967, DNSMOS 3.12 / 2.33 (original 0.0 % / 0.0 %, 0.972 / 0.964, 2.93 / 2.28) |
| bf16 | [`bf16/special_tokens_map.json`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/bf16/special_tokens_map.json) | 1.6 kB | `068594063e37` | WER 0.0 % / CER 0.0 %, speaker sim 0.973 / 0.967, DNSMOS 3.12 / 2.33 (original 0.0 % / 0.0 %, 0.972 / 0.964, 2.93 / 2.28) |
| bf16 | [`bf16/tokenizer.json`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/bf16/tokenizer.json) | 3.7 MB | `f8984687e4a9` | WER 0.0 % / CER 0.0 %, speaker sim 0.973 / 0.967, DNSMOS 3.12 / 2.33 (original 0.0 % / 0.0 %, 0.972 / 0.964, 2.93 / 2.28) |
| bf16 | [`bf16/tokenizer_config.json`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/bf16/tokenizer_config.json) | 5.0 kB | `418be464e38f` | WER 0.0 % / CER 0.0 %, speaker sim 0.973 / 0.967, DNSMOS 3.12 / 2.33 (original 0.0 % / 0.0 %, 0.972 / 0.964, 2.93 / 2.28) |
| fp16 | [`fp16/config.json`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/fp16/config.json) | 6.6 kB | `0b5bee352207` | WER 0.0 % / CER 0.0 %, speaker sim 0.970 / 0.967, DNSMOS 3.06 / 2.11 |
| fp16 | [`fp16/model.safetensors`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/fp16/model.safetensors) | 4.77 GB | `88b5a05243e9` | WER 0.0 % / CER 0.0 %, speaker sim 0.970 / 0.967, DNSMOS 3.06 / 2.11 |
| fp16 | [`fp16/model.safetensors.index.json`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/fp16/model.safetensors.index.json) | 65.7 kB | `1a4680fb582f` | WER 0.0 % / CER 0.0 %, speaker sim 0.970 / 0.967, DNSMOS 3.06 / 2.11 |
| fp16 | [`fp16/special_tokens_map.json`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/fp16/special_tokens_map.json) | 1.6 kB | `068594063e37` | WER 0.0 % / CER 0.0 %, speaker sim 0.970 / 0.967, DNSMOS 3.06 / 2.11 |
| fp16 | [`fp16/tokenizer.json`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/fp16/tokenizer.json) | 3.7 MB | `f8984687e4a9` | WER 0.0 % / CER 0.0 %, speaker sim 0.970 / 0.967, DNSMOS 3.06 / 2.11 |
| fp16 | [`fp16/tokenizer_config.json`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/fp16/tokenizer_config.json) | 5.0 kB | `418be464e38f` | WER 0.0 % / CER 0.0 %, speaker sim 0.970 / 0.967, DNSMOS 3.06 / 2.11 |
| 8bit | [`8bit/config.json`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/8bit/config.json) | 6.8 kB | `775a1f6b74b1` | WER 0.0 % / CER 0.0 %, speaker sim 0.975 / 0.966, DNSMOS 3.12 / 2.31 |
| 8bit | [`8bit/model.safetensors`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/8bit/model.safetensors) | 2.62 GB | `818a3bb5564d` | WER 0.0 % / CER 0.0 %, speaker sim 0.975 / 0.966, DNSMOS 3.12 / 2.31 |
| 8bit | [`8bit/model.safetensors.index.json`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/8bit/model.safetensors.index.json) | 0.1 MB | `1f3b958dfdf9` | WER 0.0 % / CER 0.0 %, speaker sim 0.975 / 0.966, DNSMOS 3.12 / 2.31 |
| 8bit | [`8bit/special_tokens_map.json`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/8bit/special_tokens_map.json) | 1.6 kB | `068594063e37` | WER 0.0 % / CER 0.0 %, speaker sim 0.975 / 0.966, DNSMOS 3.12 / 2.31 |
| 8bit | [`8bit/tokenizer.json`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/8bit/tokenizer.json) | 3.7 MB | `f8984687e4a9` | WER 0.0 % / CER 0.0 %, speaker sim 0.975 / 0.966, DNSMOS 3.12 / 2.31 |
| 8bit | [`8bit/tokenizer_config.json`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/8bit/tokenizer_config.json) | 5.0 kB | `418be464e38f` | WER 0.0 % / CER 0.0 %, speaker sim 0.975 / 0.966, DNSMOS 3.12 / 2.31 |
| 6bit | [`6bit/config.json`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/6bit/config.json) | 6.8 kB | `13c7d8db7354` | WER 0.0 % / CER 0.0 %, speaker sim 0.973 / 0.966, DNSMOS 3.07 / 2.55 |
| 6bit | [`6bit/model.safetensors`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/6bit/model.safetensors) | 2.05 GB | `90c38a61655b` | WER 0.0 % / CER 0.0 %, speaker sim 0.973 / 0.966, DNSMOS 3.07 / 2.55 |
| 6bit | [`6bit/model.safetensors.index.json`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/6bit/model.safetensors.index.json) | 0.1 MB | `312445674ab9` | WER 0.0 % / CER 0.0 %, speaker sim 0.973 / 0.966, DNSMOS 3.07 / 2.55 |
| 6bit | [`6bit/special_tokens_map.json`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/6bit/special_tokens_map.json) | 1.6 kB | `068594063e37` | WER 0.0 % / CER 0.0 %, speaker sim 0.973 / 0.966, DNSMOS 3.07 / 2.55 |
| 6bit | [`6bit/tokenizer.json`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/6bit/tokenizer.json) | 3.7 MB | `f8984687e4a9` | WER 0.0 % / CER 0.0 %, speaker sim 0.973 / 0.966, DNSMOS 3.07 / 2.55 |
| 6bit | [`6bit/tokenizer_config.json`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/6bit/tokenizer_config.json) | 5.0 kB | `418be464e38f` | WER 0.0 % / CER 0.0 %, speaker sim 0.973 / 0.966, DNSMOS 3.07 / 2.55 |
| 4bit | [`4bit/config.json`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/4bit/config.json) | 6.8 kB | `096014d84678` | WER 0.0 % / CER 0.0 %, speaker sim 0.972 / 0.964, DNSMOS 3.12 / 2.22 |
| 4bit | [`4bit/model.safetensors`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/4bit/model.safetensors) | 1.48 GB | `54ad41f92a33` | WER 0.0 % / CER 0.0 %, speaker sim 0.972 / 0.964, DNSMOS 3.12 / 2.22 |
| 4bit | [`4bit/model.safetensors.index.json`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/4bit/model.safetensors.index.json) | 0.1 MB | `90e318b263fd` | WER 0.0 % / CER 0.0 %, speaker sim 0.972 / 0.964, DNSMOS 3.12 / 2.22 |
| 4bit | [`4bit/special_tokens_map.json`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/4bit/special_tokens_map.json) | 1.6 kB | `068594063e37` | WER 0.0 % / CER 0.0 %, speaker sim 0.972 / 0.964, DNSMOS 3.12 / 2.22 |
| 4bit | [`4bit/tokenizer.json`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/4bit/tokenizer.json) | 3.7 MB | `f8984687e4a9` | WER 0.0 % / CER 0.0 %, speaker sim 0.972 / 0.964, DNSMOS 3.12 / 2.22 |
| 4bit | [`4bit/tokenizer_config.json`](https://huggingface.co/moonbox-dev/VoxCPM2-MLX/blob/main/4bit/tokenizer_config.json) | 5.0 kB | `418be464e38f` | WER 0.0 % / CER 0.0 %, speaker sim 0.972 / 0.964, DNSMOS 3.12 / 2.22 |
## How to run
VoxCPM2 (2B, tokenizer-free multilingual TTS with 48 kHz output: zero-shot speech, voice design and voice cloning)
for [mlx-audio](https://github.com/Blaizzy/mlx-audio). Each folder is a complete mlx-audio model: `model.safetensors`
(the MiniCPM backbone, residual LM, local encoder, flow-matching DiT and the AudioVAE), `config.json` and the
tokenizer. Tested with `mlx-audio==0.6.5`:
```python
from mlx_audio.tts.utils import load_model
import numpy as np, soundfile as sf
model = load_model("8bit") # a downloaded folder
res = list(model.generate(text="Hello from Moonbox.",
ref_audio="voice.wav", # voice cloning
prompt_audio="voice.wav", prompt_text="Exact transcript of voice.wav."))
sf.write("out.wav", np.concatenate([np.array(r.audio) for r in res]), res[0].sample_rate) # 48 kHz
```
Voice design: `model.generate(text=..., instruct="A young woman, warm and gentle voice")`. Defaults are the
original's: 10 flow-matching steps, CFG 2.0.
The tokenizer is the plain Llama tokenizer: mlx-audio splits multi-character Chinese tokens itself, which the
original repository does with remote code (`tokenization_voxcpm2.py`, left out here). `8bit/`, `6bit/` and `4bit/`
quantize the transformer layers (MLX affine, group size 64); the AudioVAE's convolutions stay bf16.
**How they were checked.** Voice cloning of 6 English and 4 Chinese sentences, two seeds each (the prompt both as
reference audio and as continuation prefix), with the original PyTorch model (float32, `voxcpm` 2.0.3, its bad-case
retry off) and every MLX build. Judged like the original: Whisper large-v3-turbo WER (English) / CER (Chinese),
WavLM-base-plus-sv speaker similarity to the prompt, DNSMOS P.835. Pass: error rate at most 2 points above the
original's, similarity at most 0.03 below, DNSMOS at most 0.2 below. The Chinese prompt is a noisy AISHELL clip,
which the model reproduces (hence the low Chinese DNSMOS for every system).
| Precision | Size | WER en / CER zh | Speaker similarity en / zh | DNSMOS en / zh | Real-time factor, M5 24 GB |
|---|---|---|---|---|---|
| original (PyTorch fp32, MPS) | - | 0.0 % / 0.0 % | 0.972 / 0.964 | 2.93 / 2.28 | 1.89 |
| `bf16/` | 4.77 GB | 0.0 % / 0.0 % | 0.973 / 0.967 | 3.12 / 2.33 | 1.08 |
| `fp16/` | 4.77 GB | 0.0 % / 0.0 % | 0.970 / 0.967 | 3.06 / 2.11 | 1.12 |
| `8bit/` | 2.62 GB | 0.0 % / 0.0 % | 0.975 / 0.966 | 3.12 / 2.31 | 1.29 |
| `6bit/` | 2.05 GB | 0.0 % / 0.0 % | 0.973 / 0.966 | 3.07 / 2.55 | 1.30 |
| `4bit/` | 1.48 GB | 0.0 % / 0.0 % | 0.972 / 0.964 | 3.12 / 2.22 | 1.14 |
Real-time factor = generation time / audio length (lower is faster), one sentence at a time.
## How these were made
From [VoxCPM2](https://huggingface.co/openbmb/VoxCPM2) at revision `7936080b41`: the original weights loaded in MLX and stored as bf16 or fp16, or
quantized with MLX's affine quantization (`8bit`, `6bit`, `4bit`, group size 64; layers MLX can't quantize, such as
convolutions and norms, stay bf16). Every build was checked against the original before upload; the `sha256` column
above is what the files must hash to.
## License
Apache-2.0. These files are a derivative of the original model and carry its license: it is copied in this
repository, and it governs your use of these files. The original model card follows unchanged.
---
# Original model card
# VoxCPM2
**VoxCPM2** is a tokenizer-free, diffusion autoregressive Text-to-Speech model β€” **2B parameters**, **30 languages**, **48kHz** audio output, trained on over **2 million hours** of multilingual speech data.
[![GitHub](https://img.shields.io/badge/GitHub-VoxCPM-blue?logo=github)](https://github.com/OpenBMB/VoxCPM)
[![Docs](https://img.shields.io/badge/Docs-ReadTheDocs-8CA1AF)](https://voxcpm.readthedocs.io/en/latest/)
[![Demo](https://img.shields.io/badge/Live%20Playground-Demo-orange)](https://huggingface.co/spaces/OpenBMB/VoxCPM-Demo)
[![Audio Samples](https://img.shields.io/badge/Audio%20Samples-Demo%20Page-green)](https://openbmb.github.io/voxcpm2-demopage)
[![Discord](https://img.shields.io/badge/Discord-VoxCPM-5865F2?logo=discord&logoColor=white)](https://discord.gg/KZUx7tVNwz)
[![Lark](https://img.shields.io/badge/飞书羀-VoxCPM-00D6B9?logo=lark&logoColor=white)](https://applink.feishu.cn/client/chat/chatter/add_by_link?link_token=acds0b9d-23d8-4d7e-b696-d200f3e22a7f) [![MiniCPM Wiki](https://img.shields.io/badge/MiniCPM-Wiki-1677FF?logo=readthedocs&logoColor=white)](https://openbmb.cn/docs/)
## Highlights
- 🌍 **30-Language Multilingual** β€” No language tag needed; input text in any supported language directly
- 🎨 **Voice Design** β€” Generate a novel voice from a natural-language description alone (gender, age, tone, emotion, pace…); no reference audio required
- πŸŽ›οΈ **Controllable Cloning** β€” Clone any voice from a short clip, with optional style guidance to steer emotion, pace, and expression while preserving timbre
- πŸŽ™οΈ **Ultimate Cloning** β€” Provide reference audio + its transcript for audio-continuation cloning; every vocal nuance faithfully reproduced
- πŸ”Š **48kHz Studio-Quality Output** β€” Accepts 16kHz reference; outputs 48kHz via AudioVAE V2's built-in super-resolution, no external upsampler needed
- 🧠 **Context-Aware Synthesis** β€” Automatically infers appropriate prosody and expressiveness from text content
- ⚑ **Real-Time Streaming** β€” RTF as low as ~0.3 on NVIDIA RTX 4090, and ~0.13 accelerated by [Nano-VLLM](https://github.com/a710128/nanovllm-voxcpm)
- πŸ“œ **Fully Open-Source & Commercial-Ready** β€” Apache-2.0 license, free for commercial use
<summary><b>Supported Languages (30)</b></summary>
Arabic, Burmese, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Indonesian, Italian, Japanese, Khmer, Korean, Lao, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swahili, Swedish, Tagalog, Thai, Turkish, Vietnamese
Chinese Dialects: 四川话, η²€θ―­, 吴语, δΈœεŒ—θ―, 河南话, ι™•θ₯Ώθ―, 山东话, 倩ζ΄₯话, 闽南话
## Quick Start
### Installation
```bash
pip install voxcpm
```
**Requirements:** Python β‰₯ 3.10, PyTorch β‰₯ 2.5.0, CUDA β‰₯ 12.0 Β· [Full Quick Start β†’](https://voxcpm.readthedocs.io/en/latest/quickstart.html)
### Text-to-Speech
```python
from voxcpm import VoxCPM
import soundfile as sf
model = VoxCPM.from_pretrained("openbmb/VoxCPM2", load_denoiser=False)
wav = model.generate(
text="VoxCPM2 brings multilingual support, creative voice design, and controllable voice cloning.",
cfg_value=2.0,
inference_timesteps=10,
)
sf.write("output.wav", wav, model.tts_model.sample_rate)
```
### Voice Design
Put the voice description in parentheses at the start of `text`, followed by the content to synthesize:
```python
wav = model.generate(
text="(A young woman, gentle and sweet voice)Hello, welcome to VoxCPM2!",
cfg_value=2.0,
inference_timesteps=10,
)
sf.write("voice_design.wav", wav, model.tts_model.sample_rate)
```
### Controllable Voice Cloning
```python
# Basic cloning
wav = model.generate(
text="This is a cloned voice generated by VoxCPM2.",
reference_wav_path="speaker.wav",
)
sf.write("clone.wav", wav, model.tts_model.sample_rate)
# Cloning with style control
wav = model.generate(
text="(slightly faster, cheerful tone)This is a cloned voice with style control.",
reference_wav_path="speaker.wav",
cfg_value=2.0,
inference_timesteps=10,
)
sf.write("controllable_clone.wav", wav, model.tts_model.sample_rate)
```
### Ultimate Cloning
Provide both the reference audio and its exact transcript for maximum fidelity. Pass the same clip to both `reference_wav_path` and `prompt_wav_path` for highest similarity:
```python
wav = model.generate(
text="This is an ultimate cloning demonstration using VoxCPM2.",
prompt_wav_path="speaker_reference.wav",
prompt_text="The transcript of the reference audio.",
reference_wav_path="speaker_reference.wav",
)
sf.write("hifi_clone.wav", wav, model.tts_model.sample_rate)
```
### Streaming
```python
import numpy as np
chunks = []
for chunk in model.generate_streaming(text="Streaming is easy with VoxCPM!"):
chunks.append(chunk)
wav = np.concatenate(chunks)
sf.write("streaming.wav", wav, model.tts_model.sample_rate)
```
## Model Details
| Property | Value |
|---|---|
| Architecture | Tokenizer-free Diffusion Autoregressive (LocEnc β†’ TSLM β†’ RALM β†’ LocDiT) |
| Backbone | Based on MiniCPM-4, totally 2B parameters |
| Audio VAE | AudioVAE V2 (asymmetric encode/decode, 16kHz in β†’ 48kHz out) |
| Training Data | 2M+ hours multilingual speech |
| LM Token Rate | 6.25 Hz |
| Max Sequence Length | 8192 tokens |
| dtype | bfloat16 |
| VRAM | ~8 GB |
| RTF (RTX 4090) | ~0.30 (standard) / ~0.13 (Nano-vLLM) |
## Performance
VoxCPM2 achieves state-of-the-art or competitive results on major zero-shot and controllable TTS benchmarks.
See the [GitHub repo](https://github.com/OpenBMB/VoxCPM#-performance) for full benchmark tables (Seed-TTS-eval, CV3-eval, InstructTTSEval, MiniMax Multilingual Test).
## Fine-tuning
VoxCPM2 supports both full SFT and LoRA fine-tuning with as little as 5–10 minutes of audio:
```bash
# LoRA fine-tuning (recommended)
python scripts/train_voxcpm_finetune.py \
--config_path conf/voxcpm_v2/voxcpm_finetune_lora.yaml
# Full fine-tuning
python scripts/train_voxcpm_finetune.py \
--config_path conf/voxcpm_v2/voxcpm_finetune_all.yaml
```
See the [Fine-tuning Guide](https://voxcpm.readthedocs.io/en/latest/finetuning/finetune.html) for full instructions.
## Limitations
- Voice Design and Style Control results may vary between runs; generating 1–3 times is recommended to obtain the desired output.
- Performance varies across languages depending on training data availability.
- Occasional instability may occur with very long or highly expressive inputs.
- **Strictly forbidden** to use for impersonation, fraud, or disinformation. AI-generated content should be clearly labeled.
## Citation
```bibtex
@article{voxcpm2_2026,
title = {VoxCPM2: Tokenizer-Free TTS for Multilingual Speech Generation, Creative Voice Design, and True-to-Life Cloning},
author = {VoxCPM Team},
journal = {GitHub},
year = {2026},
}
@article{voxcpm2025,
title = {VoxCPM: Tokenizer-Free TTS for Context-Aware Speech Generation and True-to-Life Voice Cloning},
author = {Zhou, Yixuan and Zeng, Guoyang and Liu, Xin and Li, Xiang and
Yu, Renjie and Wang, Ziyang and Ye, Runchuan and Sun, Weiyue and
Gui, Jiancheng and Li, Kehan and Wu, Zhiyong and Liu, Zhiyuan},
journal = {arXiv preprint arXiv:2509.24650},
year = {2025},
}
```
## License
Released under the [Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0) license, free for commercial use. For production deployments, we recommend thorough testing and safety evaluation tailored to your use case.