Deploy ModelScope FunASR Model on Arm Servers
Introduction
Introduction to Automatic Speech Recognition
ModelScope - an Open Source Pre-trained AI Models Hub
Building ASR Applications with ModelScope
Next Steps

  Log an issue

  Fork and edit

  Discuss on Discord

About this Learning Path

Skill level:  Introductory
Reading time:  1 hr
Last updated:  11 Sep 2026
Skill level:
  Introductory
Reading time:
  1 hr
Last updated:
  11 Sep 2026
Author:Odin Shen, Arm
Arm IP:Neoverse
Tags: ML AWS Graviton Microsoft Azure Cobalt Google Axion Oracle Cloud Infrastructure (OCI) Ampere Compute Linux ModelScope FunASR LLM Generative AI Python
Author:
  • Odin Shen, Arm
Arm IP:
Neoverse
Tags:
ML AWS Graviton Microsoft Azure Cobalt Google Axion Oracle Cloud Infrastructure (OCI) Ampere Compute Linux ModelScope FunASR LLM Generative AI Python

Who is this for?

This is an introductory topic for developers interested in learning how to deploy the ModelScope FunASR Chinese Automatic Speech Recognition (ASR) model on Arm-based servers.

What will you learn?

Upon completion of this Learning Path, you will be able to:

  • Leverage open-source large language models and tools to build Chinese ASR applications.
  • Deploy real-time Chinese speech recognition, punctuation restoration, and sentiment analysis using FunASR.
  • Describe how to accelerate ModelScope models on Arm-based servers for enhanced performance and efficiency.

Prerequisites

Before starting, you will need the following:

  • An Arm-based instance from a cloud service provider, or a local Arm Linux computer with at least 8 CPUs and 16GB of RAM

Summary

AI-assisted

This summary was drafted with an approved AI-assisted workflow and reviewed by Arm contributors before publication. Human technical review remains part of the process so the final page reflects engineering rigor, accuracy, and Arm editorial standards.

Close
?
You’ll deploy a Chinese speech-recognition workflow on Arm-based Linux servers with ModelScope and FunASR. First, you’ll prepare an Arm Ubuntu environment, install the pinned FunASR release, load a pretrained model, and run speech-to-text. Next, you’ll optionally enable punctuation restoration and sentiment analysis, then review the transcription and analysis output from the completed pipeline.

Frequently asked questions

AI-assisted

These FAQs were drafted with an approved AI-assisted workflow and reviewed by Arm contributors before publication. Human technical review remains part of the process so the final page reflects engineering rigor, accuracy, and Arm editorial standards.

Close
?
Which FunASR version should I use for the examples?
Use funasr==1.2.3. Results might vary with other versions.
What should I check on my server before installing anything?
Verify that you’re on an Arm-based machine running Ubuntu 22.04 LTS or later with at least 8 cores, 16GB RAM, and 30GB of free disk space.
What result should I expect when the ASR pipeline runs successfully?
You should see Chinese speech transcribed to text, with optional punctuation restoration and sentiment analysis outputs.
Where do the models used in the examples come from?
The models are pre-trained and come from ModelScope, an open-source platform designed to simplify the integration of AI models into applications.
Which Python version does the optimized PyTorch setup require?
Use Python 3.10. If your current version is lower or higher, install python3.10, configure the python3 alternatives, and confirm the active version with python --version.
Next