Overview

MP SENet DNS is a specialized audio-to-audio model designed for high-fidelity deep noise suppression. Unlike generic noise gates, it leverages a Squeeze-and-Excitation Network architecture to dynamically weight channel features, allowing it to isolate target speech from complex background interference with minimal spectral distortion. For developers building VoIP applications, accessibility tools, or podcasting software, this model provides a computationally efficient way to implement real-time audio cleanup. It integrates easily into signal processing pipelines and offers a significant performance boost in signal-to-noise ratio (SNR) compared to traditional subtractive methods, making it ideal for deployment in environments with unpredictable acoustic noise.

Highlights

Efficient deep noise suppression via SE-Net architecture
High-fidelity speech recovery with minimal audio artifacts
Ideal for real-time VoIP and communication tools
Permissive MIT license for flexible commercial integration

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("JacobLinCool/MP-SENet-DNS")
tokenizer = AutoTokenizer.from_pretrained("JacobLinCool/MP-SENet-DNS")

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 JacobLinCool/MP-SENet-DNS

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

huggingface-cli download JacobLinCool/MP-SENet-DNS 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('JacobLinCool/MP-SENet-DNS')

Git Download

Make sure git-lfs is installed first

Git Download

git lfs install
git clone https://huggingface.co/JacobLinCool/MP-SENet-DNS

To skip LFS large-file downloads, use:

Skip LFS

GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/JacobLinCool/MP-SENet-DNS

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('JacobLinCool/MP-SENet-DNS')
tokenizer = AutoTokenizer.from_pretrained('JacobLinCool/MP-SENet-DNS')

Full Documentation

来源: HuggingFace

---
license: mit
pipeline_tag: audio-to-audio
tags:

arxiv:2308.08926

audio

denoising

model_hub_mixin

pytorch_model_hub_mixin

speech

speech-enhancement

---

This model has been pushed to the Hub using the PytorchModelHubMixin integration:

Library: https://github.com/yxlu-0102/MP-SENet

Docs: [More Information Needed]