silero vad onnx

Provideristupakov
Categoryvoice-activity-detection
Licensemit
Downloads5.0K
Stars0

Overview

Silero VAD is a high-performance voice activity detection model optimized for production environments via the ONNX runtime. Unlike traditional energy-based VADs, it leverages a deep learning architecture to accurately distinguish human speech from background noise with minimal latency. For developers, this means a lightweight, dependency-free integration that works across various platforms without requiring a full PyTorch installation. It is ideal for preprocessing pipelines in STT (Speech-to-Text) systems, reducing API costs by filtering silence, and implementing wake-word triggers. Compared to heavier models, Silero offers a superior balance of accuracy and computational efficiency, making it suitable for both server-side scaling and edge deployment.

Highlights

  • ONNX runtime compatibility for cross-platform deployment
  • Low-latency speech detection with minimal CPU overhead
  • Highly accurate noise rejection compared to energy-based VAD
  • MIT licensed for flexible commercial integration
  • Optimized for preprocessing in STT and VoIP pipelines

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("istupakov/silero-vad-onnx")
tokenizer = AutoTokenizer.from_pretrained("istupakov/silero-vad-onnx")

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

Download the full repository 
huggingface-cli download istupakov/silero-vad-onnx

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

Download a single file to a local folder (e.g. config.json into ./dir) 
huggingface-cli download istupakov/silero-vad-onnx config.json --local-dir ./dir

See the official docs for more CLI options

SDK Download

SDK Download 
# 模型下载
from huggingface_hub import snapshot_download
model_dir = snapshot_download('istupakov/silero-vad-onnx')

Git Download

Make sure git-lfs is installed first

Git Download 
git lfs install
git clone https://huggingface.co/istupakov/silero-vad-onnx

To skip LFS large-file downloads, use:

Skip LFS 
GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/istupakov/silero-vad-onnx

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

Install Transformers 
pip install -U transformers torch

Load the model and run inference

Load the model and run inference 
from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained('istupakov/silero-vad-onnx')
tokenizer = AutoTokenizer.from_pretrained('istupakov/silero-vad-onnx')

Full Documentation

来源: HuggingFace

---
license: mit
pipeline_tag: voice-activity-detection
tags:
- voice-activity-detection
- vad
- onnx
- onnx-asr
---
Silero VAD model (version 6.2) in ONNX format for onnx-asr.

Install onnx-asr

shell
pip install onnx-asr[cpu,hub]

Load Parakeet v3 ASR with Silero VAD and recognize wav file

py
import onnx_asr
vad = onnx_asr.load_vad("silero")
model = onnx_asr.load_model("nemo-parakeet-tdt-0.6b-v3").with_vad(vad)
for res in model.recognize("test.wav"):
    print(res)

View official source
Join our Telegram