Name: segmentation 3.0
Author: pyannote

Overview

Segmentation 3.0 by pyannote is a specialized voice activity detection (VAD) model designed to distinguish human speech from background noise with high precision. For developers building speech-to-text pipelines or conversational AI, this model serves as a critical preprocessing layer to reduce noise-induced transcription errors and optimize compute by stripping silence. It integrates seamlessly into Python-based audio workflows, offering a lightweight alternative to full-scale diarization systems when only speech boundary detection is required. Compared to generic VADs, it provides more robust performance across varying acoustic environments, making it ideal for cleaning raw audio streams before they hit an ASR engine.

Highlights

High-precision voice activity detection for audio preprocessing
Reduces ASR errors by filtering non-speech segments
MIT licensed for flexible commercial and open-source integration
Optimized for Python-based speech processing pipelines
Efficiently separates human speech from ambient background noise

Usage

Install

# Install Hugging Face transformers
pip install transformers torch

SDK Usage

# Load model with transformers
from transformers import AutoModel, AutoTokenizer

model = AutoModel.from_pretrained("pyannote/segmentation-3.0")
tokenizer = AutoTokenizer.from_pretrained("pyannote/segmentation-3.0")

Hugging Face Download

We recommend downloading the model via the Hugging Face CLI or Hub SDK.

Guidance：Before downloading, install huggingface_hub with:

Guidance

pip install -U huggingface_hub

CLI Download

Download the full repository

huggingface-cli download pyannote/segmentation-3.0

Download a single file to a local folder (e.g. config.json into ./dir)

huggingface-cli download pyannote/segmentation-3.0 config.json --local-dir ./dir

See the official docs for more CLI options

SDK Download

# 模型下载
from huggingface_hub import snapshot_download
model_dir = snapshot_download('pyannote/segmentation-3.0')

Git Download

Make sure git-lfs is installed first

Git Download

git lfs install
git clone https://huggingface.co/pyannote/segmentation-3.0

To skip LFS large-file downloads, use:

Skip LFS

GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/pyannote/segmentation-3.0

Model files are hosted on the Hugging Face Hub — download directly via HF CLI / SDK / Git, not through this site.

PyTorch / Transformers Usage

Install Transformers

pip install -U transformers torch

Load the model and run inference

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained('pyannote/segmentation-3.0')
tokenizer = AutoTokenizer.from_pretrained('pyannote/segmentation-3.0')

Model Download

We recommend downloading the model via the ModelScope CLI or SDK.

Guidance：Before downloading, install ModelScope with:

Guidance

pip install modelscope

CLI Download

Download the full repository

modelscope download --model pyannote/segmentation-3.0

Download a single file to a local folder (e.g. README.md into ./dir)

modelscope download --model pyannote/segmentation-3.0 README.md --local_dir ./dir

See the docs for more CLI options

SDK Download

# 模型下载
from modelscope import snapshot_download
model_dir = snapshot_download('pyannote/segmentation-3.0')

Git Download

Make sure git-lfs is installed first

Git Download

git lfs install
git clone https://www.modelscope.cn/pyannote/segmentation-3.0.git

To skip LFS large-file downloads, use:

Skip LFS

GIT_LFS_SKIP_SMUDGE=1 git clone https://www.modelscope.cn/pyannote/segmentation-3.0.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 Quickstart

Install the ModelScope library

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

Load the model and run inference

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

p = pipeline('text-generation', 'pyannote/segmentation-3.0')