Skip to content
AI Tutorials
Text-to-Speech
CuteTTS

Install OPPO CuteTTS TTS with Conda

Introduction

CuteTTS: a lightweight continuous autoregressive TTS model that runs efficiently on GPUs, CPUs, and Apple silicon.

GitHub: https://github.com/OPPO-Mente-Lab/CuteTTS
Hugging Face: https://hf.co/OPPOer/CuteTTS
Hugging Face: https://hf.co/OPPOer/CuteTTS-distill
Demo: https://hf.co/spaces/hugging-apps/cutetts-demo
Paper: https://arxiv.org/abs/2608.08638

CuteTTS

  • A lightweight (~230M-parameter) continuous autoregressive TTS model that runs efficiently on GPUs, CPUs, and Apple silicon.
  • Ultra-low latency: ~40 ms to the first audio chunk and a throughput of ~9× real time on an NVIDIA RTX 4090.
  • Excellent speech quality and voice cloning performance.
  • Web demo, Python API, and CLI.
  • Multilingual support: English, Chinese, French, German, and Spanish.

Prerequisites

System requirements:

  • Operating System: Windows 10/11 (64-bit), macOS, or Linux (Debian/Ubuntu).
  • Python: version >= 3.10 required
  • Disk Space: 10GB+ recommended (for dependencies and model cache). At least 400 MB for Miniconda; 3 GB+ for full Anaconda.
  • The GPU is optional but HIGHLY Recommended for Performance.
  • Internet: For downloading dependencies and models from Hugging Face Hub.
Requirements CPU RAM
Minimum Intel Core i3 or AMD Ryzen 3 8 GB
Recommended Intel Core i5 or AMD Ryzen 5 16 GB
High-performance Intel Core i7/i9 or AMD Ryzen 7/9 32 GB
Environment Run this Command
CPU only pip3 install torch torchvision
Apple Silicon pip3 install torch torchvision
AMD ROCm Install PyTorch for ROCm
CUDA 11.8 pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118
CUDA 12.1 pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121
CUDA 12.6 pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu126
CUDA 12.8 pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu128
CUDA 13.0 pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu130

Video tutorial

Coming soon!

Step 1. Install Miniconda Package

Step 2. Create Conda Environment

Create an environment.yml file

name: cutetts
channels:
- conda-forge
- defaults

dependencies:
# Python version (requires >= 3.10)
- python=3.12

- pip
- pip:
    # PyTorch CPU only
    - torch
    - torchaudio

    # Install CuteTTS directly from GitHub.
    - git+https://github.com/OPPO-Mente-Lab/CuteTTS.git

    # Download the model from Hugging Face / ModelScope
    - huggingface_hub

Activate conda environment:

conda env create -f environment.yml
conda activate cutetts

Clone and install CuteTTS from source (run after activating the environment)

git clone https://github.com/OPPO-Mente-Lab/CuteTTS.git
cd CuteTTS

Download weights

Download CuteTTS or CuteTTS-distill into ./model:

# Create the model directory
mkdir model

# Download the CuteTTS model (or CuteTTS-distill for lower-end hardware)
hf download OPPOer/CuteTTS --local-dir ./model/CuteTTS
hf download OPPOer/CuteTTS-distill --local-dir ./model/CuteTTS-distill

Directory structure

The expected directory structure is:

D:\cute\
├── CuteTTS\
│   ├── assets\
│   ├── LICENSES\
│   ├── model\
│   │   ├── CuteTTS\
│   │   └── CuteTTS-distill\
│   ├── src\
│   │   └── cutetts\
│   ├── .gitignore
│   ├── LICENSE
│   ├── NOTICE
│   ├── README.md
│   ├── README_zh.md
│   ├── infer.py
│   └── pyproject.toml
├── cute_test.py
└── environment.yml

Step 3. Run the Inference

Command line 

# sanity check
cutetts --help
usage: cutetts [-h] --model-dir MODEL_DIR --text TEXT [--mode {tts,voice_clone}] [--reference-audio REFERENCE_AUDIO] [--output OUTPUT] [--device DEVICE] [--cfg-strength CFG_STRENGTH]
               [--diffusion-steps DIFFUSION_STEPS] [--diffusion-sway-coefficient DIFFUSION_SWAY_COEFFICIENT] [--max-decode-length MAX_DECODE_LENGTH] [--seed SEED] [--no-progress]

Generate speech with CuteTTS.

options:
  -h, --help            show this help message and exit
  --model-dir MODEL_DIR
  --text TEXT
  --mode {tts,voice_clone}
  --reference-audio REFERENCE_AUDIO
  --output OUTPUT
  --device DEVICE       Inference device: auto, cuda, mps, or cpu (default: auto).
  --cfg-strength CFG_STRENGTH
  --diffusion-steps DIFFUSION_STEPS
  --diffusion-sway-coefficient DIFFUSION_SWAY_COEFFICIENT
  --max-decode-length MAX_DECODE_LENGTH
  --seed SEED
  --no-progress

Now, run the basic example script to synthesize speech:

cutetts --model-dir ./model/CuteTTS --mode tts --text "The voice generated by this model sounds amazing!" --output tts.wav

The result will be the audio file tts.wav

Test Voice Cloning

cutetts --model-dir ./model/CuteTTS --mode voice_clone --reference-audio ./assets/default_reference.wav --text "The voice generated by this model sounds amazing!" --output clone.wav

Python Script

Create a file named cute_test.py:

from cutetts import CuteTTS
import soundfile as sf

model = CuteTTS.from_pretrained("./CuteTTS/model/CuteTTS", device="auto")
result = model.generate("The voice generated by this model sounds amazing!", mode="tts")
sf.write("tts.wav", result.waveform.squeeze(0).numpy(), result.sample_rate)

clone = model.generate(
    "The voice generated by this model sounds amazing!",
    mode="voice_clone",
    reference_audio="./CuteTTS/assets/default_reference.wav",
)
sf.write("clone.wav", clone.waveform.squeeze(0).numpy(), clone.sample_rate)

Run the script to clone voice:

python cute_test.py

The result will be the audio file clone.wav

Web demo

Try CuteTTS without coding:

cutetts-demo --model-dir ./model --device auto --host 127.0.0.1 --port 7860

Open your browser and navigate to http://127.0.0.1:7860. The system will automatically download the required model weights from HuggingFace during this first run.

ChatterboxDia