Name: metricgan plus voicebank
Author: speechbrain

Overview

MetricGAN+ VoiceBank is a specialized audio-to-audio framework designed for speech enhancement and quality improvement. Unlike general-purpose generative models, it focuses on objective speech quality metrics to minimize the distance between degraded audio and clean references. For developers, this means a tool capable of reducing background noise and distortion while maintaining the natural characteristics of the original speaker's voice. It is particularly effective for preprocessing pipelines in ASR (Automatic Speech Recognition) or TTS (Text-to-Speech) systems where signal clarity directly impacts downstream accuracy. Built on the SpeechBrain toolkit and released under the Apache-2.0 license, it integrates easily into PyTorch workflows, offering a predictable, research-backed approach to audio restoration without the overhead of massive parameter counts.

Highlights

Optimizes speech quality using objective metric-based loss functions.
Reduces noise and distortion for cleaner audio signals.
Seamless integration with PyTorch via the SpeechBrain toolkit.
Apache-2.0 license allows for flexible commercial deployment.
Ideal for enhancing ASR and TTS input data.

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("speechbrain/metricgan-plus-voicebank")
tokenizer = AutoTokenizer.from_pretrained("speechbrain/metricgan-plus-voicebank")

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 speechbrain/metricgan-plus-voicebank

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

huggingface-cli download speechbrain/metricgan-plus-voicebank 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('speechbrain/metricgan-plus-voicebank')

Git Download

Make sure git-lfs is installed first

Git Download

git lfs install
git clone https://huggingface.co/speechbrain/metricgan-plus-voicebank

To skip LFS large-file downloads, use:

Skip LFS

GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/speechbrain/metricgan-plus-voicebank

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('speechbrain/metricgan-plus-voicebank')
tokenizer = AutoTokenizer.from_pretrained('speechbrain/metricgan-plus-voicebank')

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 speechbrain/metricgan-plus-voicebank

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

modelscope download --model speechbrain/metricgan-plus-voicebank 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('speechbrain/metricgan-plus-voicebank')

Git Download

Make sure git-lfs is installed first

Git Download

git lfs install
git clone https://www.modelscope.cn/speechbrain/metricgan-plus-voicebank.git

To skip LFS large-file downloads, use:

Skip LFS

GIT_LFS_SKIP_SMUDGE=1 git clone https://www.modelscope.cn/speechbrain/metricgan-plus-voicebank.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', 'speechbrain/metricgan-plus-voicebank')

Full Documentation

来源: HuggingFace

---
language: "en"
tags:

audio-to-audio

speech-enhancement

PyTorch

speechbrain

license: "apache-2.0"
datasets:

Voicebank

DEMAND

metrics:

PESQ

STOI

inference: false
---

MetricGAN-trained model for Enhancement

This repository provides all the necessary tools to perform enhancement with
SpeechBrain. For a better experience we encourage you to learn more about
SpeechBrain. The model performance is:

| Release | Test PESQ | Test STOI |
|:-----------:|:-----:| :-----:|
| 21-04-27 | 3.15 | 93.0 |

Install SpeechBrain

First of all, please install SpeechBrain with the following command:

code

pip install speechbrain

Please notice that we encourage you to read our tutorials and learn more about
SpeechBrain.

Pretrained Usage

To use the mimic-loss-trained model for enhancement, use the following simple code:

python

import torch
import torchaudio
from speechbrain.inference.enhancement import SpectralMaskEnhancement
enhance_model = SpectralMaskEnhancement.from_hparams(
    source="speechbrain/metricgan-plus-voicebank",
    savedir="pretrained_models/metricgan-plus-voicebank",
)
Load and add fake batch dimension
noisy = enhance_model.load_audio(
    "speechbrain/metricgan-plus-voicebank/example.wav"
).unsqueeze(0)
Add relative length tensor
enhanced = enhance_model.enhance_batch(noisy, lengths=torch.tensor([1.]))
Saving enhanced signal on disk
torchaudio.save('enhanced.wav', enhanced.cpu(), 16000)

The system is trained with recordings sampled at 16kHz (single channel).
The code will automatically normalize your audio (i.e., resampling + mono channel selection) when calling *enhance_file* if needed. Make sure your input tensor is compliant with the expected sampling rate if you use *enhance_batch* as in the example.

Inference on GPU

To perform inference on the GPU, add run_opts={"device":"cuda"} when calling the from_hparams method.

Training

The model was trained with SpeechBrain (d0accc8). To train it from scratch follows these steps: 1. Clone SpeechBrain:

bash

git clone https://github.com/speechbrain/speechbrain/

2. Install it:

code

cd speechbrain
pip install -r requirements.txt
pip install -e .

3. Run Training:

code

cd  recipes/Voicebank/enhance/MetricGAN
python train.py hparams/train.yaml --data_folder=your_data_folder

You can find our training results (models, logs, etc) here.

Limitations

The SpeechBrain team does not provide any warranty on the performance achieved by this model when used on other datasets.

Referencing MetricGAN+

If you find MetricGAN+ useful, please cite:

code

@article{fu2021metricgan+,
  title={MetricGAN+: An Improved Version of MetricGAN for Speech Enhancement},
  author={Fu, Szu-Wei and Yu, Cheng and Hsieh, Tsun-An and Plantinga, Peter and Ravanelli, Mirco and Lu, Xugang and Tsao, Yu},
  journal={arXiv preprint arXiv:2104.03538},
  year={2021}
}

About SpeechBrain

Website: https://speechbrain.github.io/

Code: https://github.com/speechbrain/speechbrain/

HuggingFace: https://huggingface.co/speechbrain/

Citing SpeechBrain

Please, cite SpeechBrain if you use it for your research or business.

bibtex

@misc{speechbrain,
  title={{SpeechBrain}: A General-Purpose Speech Toolkit},
  author={Mirco Ravanelli and Titouan Parcollet and Peter Plantinga and Aku Rouhe and Samuele Cornell and Loren Lugosch and Cem Subakan and Nauman Dawalatabad and Abdelwahab Heba and Jianyuan Zhong and Ju-Chieh Chou and Sung-Lin Yeh and Szu-Wei Fu and Chien-Feng Liao and Elena Rastorgueva and François Grondin and William Aris and Hwidong Na and Yan Gao and Renato De Mori and Yoshua Bengio},
  year={2021},
  eprint={2106.04624},
  archivePrefix={arXiv},
  primaryClass={eess.AS},
  note={arXiv:2106.04624}
}