Name: Lightweight Speech Denoising.axera
Author: AXERA-TECH

Overview

Lightweight Speech Denoising by AXERA-TECH is a specialized speech enhancement model designed for deployment in resource-constrained environments. Unlike heavy transformer-based architectures, this model prioritizes low latency and a small memory footprint, making it ideal for integration into edge devices, mobile apps, or real-time communication pipelines. It focuses on isolating clean voice signals from background noise, providing a practical solution for developers building VoIP services, voice assistants, or accessibility tools where on-device processing is critical to reduce server costs and latency. Released under the MIT license, it offers maximum flexibility for commercial integration without restrictive licensing overhead.

Highlights

Optimized for low-latency edge device deployment
Efficient real-time background noise suppression
Permissive MIT license for commercial use
Minimal memory overhead for embedded systems

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("AXERA-TECH/Lightweight-Speech-Denoising.axera")
tokenizer = AutoTokenizer.from_pretrained("AXERA-TECH/Lightweight-Speech-Denoising.axera")

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 AXERA-TECH/Lightweight-Speech-Denoising.axera

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

huggingface-cli download AXERA-TECH/Lightweight-Speech-Denoising.axera 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('AXERA-TECH/Lightweight-Speech-Denoising.axera')

Git Download

Make sure git-lfs is installed first

Git Download

git lfs install
git clone https://huggingface.co/AXERA-TECH/Lightweight-Speech-Denoising.axera

To skip LFS large-file downloads, use:

Skip LFS

GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/AXERA-TECH/Lightweight-Speech-Denoising.axera

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('AXERA-TECH/Lightweight-Speech-Denoising.axera')
tokenizer = AutoTokenizer.from_pretrained('AXERA-TECH/Lightweight-Speech-Denoising.axera')

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 AXERA-TECH/Lightweight-Speech-Denoising.axera

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

modelscope download --model AXERA-TECH/Lightweight-Speech-Denoising.axera 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('AXERA-TECH/Lightweight-Speech-Denoising.axera')

Git Download

Make sure git-lfs is installed first

Git Download

git lfs install
git clone https://www.modelscope.cn/AXERA-TECH/Lightweight-Speech-Denoising.axera.git

To skip LFS large-file downloads, use:

Skip LFS

GIT_LFS_SKIP_SMUDGE=1 git clone https://www.modelscope.cn/AXERA-TECH/Lightweight-Speech-Denoising.axera.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', 'AXERA-TECH/Lightweight-Speech-Denoising.axera')

Full Documentation

来源: HuggingFace

---
license: mit
language:

zh

en

pipeline_tag: audio-to-audio
tags:

speech-enhancement

---

Lightweight-Speech-Denoising

This project provides lightweight speech enhancement (denoising) models optimized for Axera NPU platforms, combining the DSP framework of RNNoise and the model architecture of GTCRN.

Key Features

Good Denoising Quality — Built on GTCRN and RNNoise frameworks; strong noise suppression even at very low parameter counts.

Ultra-Lightweight Models — Smallest model under 100 KB; CMM memory footprint below 150 KB.

Minimal Operators — tiny_v5 and conv_se are pure convolutional models with very few operator types; tiny_v5 supports quantization on the operator-limited AX525 platform.

End-to-End Workflow — Full pipeline. Full project with export / quantization scripts

Multi-Platform Support — Board inference validated on AX620Q, AX630C, and AX650; PTQ quantization also available for AX620L, AX637, and AX525.

Convert tools links

For those interested in model conversion, refer to:

Pulsar2 Docs — How to Convert ONNX to axmodel

Full project with export / quantization scripts

Support Platform

AX650

- M4N-Dock（爱芯派Pro） - M.2 Accelerator Card

AX630C

- 爱芯派2 - Module-LLM - LLM630 Compute Kit

AX620Q

AX620L

AX637

AX525（only for tiny_v5）

How to use

Directory layout on device:

code

root@ax650:~# tree Lightweight-Speech-Denoising.axera
Lightweight-Speech-Denoising.axera
├── README.md
├── axmodels
│   ├── ax525_tiny_v5_setrain.axmodel
│   ├── ax620E_conv_se_setrain.axmodel
│   ├── ax620E_gtcrn_setrain.axmodel
│   ├── ax620E_tiny_v5_setrain.axmodel
│   ├── ax620L_conv_se_setrain.axmodel
│   ├── ax620L_gtcrn_setrain.axmodel
│   ├── ax620L_tiny_v5_setrain.axmodel
│   ├── ax637_conv_se_setrain.axmodel
│   ├── ax637_gtcrn_setrain.axmodel
│   ├── ax637_tiny_v5_setrain.axmodel
│   ├── ax630c_conv_se_setrain.axmodel
│   ├── ax630c_gtcrn_setrain.axmodel
│   ├── ax630c_tiny_v5_setrain.axmodel
│   ├── ax650_conv_se_setrain.axmodel
│   ├── ax650_gtcrn_setrain.axmodel
│   └── ax650_tiny_v5_setrain.axmodel
├── build_ax620q
│   └── test_se_denoise_ax
├── build_ax630c
│   └── test_se_denoise_ax
├── build_ax650
│   └── test_se_denoise_ax
├── models
│   ├── conv_se_ax650_config.ini
│   ├── gtcrn_7input_ax650_config.ini
│   ├── tiny_v5_ax650_config.ini
│   └── ...
├── run_ax620q_all.sh
├── run_ax630c_all.sh
├── run_ax650_all.sh
└── test_wavs
    └── mix.wav

Download all files from this repository to the device, then run the corresponding script for your platform:

bash

# AX650
sh run_ax650_all.sh
AX630C
sh run_ax630c_all.sh
AX620Q
sh run_ax620q_all.sh

Output .wav files will be saved to output/<platform>_all/.

Inference Results

tiny_v5

| Platform | Avg Infer (ms) | RTF | Realtime Speedup |
|----------|---------------|--------|-----------------|
| AX650 | 0.160 | 0.0117 | 85.3x |
| AX630C | 0.587 | 0.0232 | 43.1x |
| AX620Q | 0.736 | 0.0332 | 30.1x |
| AX620L | TBD | TBD | TBD |
| AX637 | TBD | TBD | TBD |
| AX525 | TBD | TBD | TBD |

conv_se

| Platform | Avg Infer (ms) | RTF | Realtime Speedup |
|----------|---------------|--------|-----------------|
| AX650 | 1.963 | 0.0365 | 27.4x |
| AX630C | 7.803 | 0.1092 | 9.2x |
| AX620Q | 14.938 | 0.1970 | 5.1x |
| AX620L | TBD | TBD | TBD |
| AX637 | TBD | TBD | TBD |

GTCRN

| Platform | Avg Infer (ms) | RTF | Realtime Speedup |
|----------|---------------|--------|-----------------|
| AX650 | 2.766 | 0.1756 | 5.7x |
| AX630C | 2.835 | 0.1820 | 5.5x |
| AX620Q | 3.535 | 0.2295 | 4.4x |
| AX620L | TBD | TBD | TBD |
| AX637 | TBD | TBD | TBD |

> Test audio: mix.wav, duration 9.77s, 16kHz mono.

TODO

[ ] AX525 board inference (quantization done)

[ ] AX620L board inference (quantization done)

[ ] AX637 board inference (quantization done)

References

RNNoise — Mozilla open-source DSP + RNN noise suppression framework; STFT/iSTFT and kiss_fft implementation reused in this project.

https://github.com/xiph/rnnoise

GTCRN — Lightweight Gated Temporal Convolutional Recurrent Network for speech enhancement.

https://github.com/Xiaobin-Rong/gtcrn