In this Learning Path, you will learn how to convert the Stable Audio Open Small model to ExecuTorch (.pte) format, then build an audio generation application to run on Android or macOS. ExecuTorch is a PyTorch framework designed for on-device inference on edge and mobile devices.
Before you begin, you need a development environment with the required software:
Create a workspace directory to manage the dependencies and repositories.
Export the WORKSPACE variable to point to this directory:
mkdir my-workspace
export WORKSPACE=$PWD/my-workspace
cd $WORKSPACE
Install Python 3.10 or newer for compatibility with the required packages:
sudo apt install -y python3.10 python3.10-venv
brew install python@3.10
brew link python@3.10 --force
Verify the installation:
python3 --version
The output is similar to:
Python 3.10.19
Install CMake to automate the build process for the audio generation application:
sudo apt update
sudo apt install cmake g++ git
brew install cmake
Verify the installation:
cmake --version
The output is similar to:
cmake version 4.2.1
See the CMake install guide for additional help.
Clone the Arm ML examples repository, which contains the scripts and application code:
cd $WORKSPACE
git clone https://github.com/Arm-Examples/ML-examples.git
cd ML-examples/kleidiai-examples/audiogen-et/
Your development environment is ready with Python 3.10, CMake, and the ML examples repository. In the next section, you’ll download the Stable Audio Open Small model from Hugging Face.