Build ONNX file in ten-vad repo.

Repo: https://github.com/TEN-framework/ten-vad#usage-onnx

Tested on Ubuntu 24.04.2 LTS on x64 CPU.

1. Clone ten-vad repo

cd
git clone https://github.com/TEN-framework/ten-vad.git

2. Download and extract onnxruntime v1.17.1

A newer version for download is in Next Steps.

sudo apt update
sudo apt upgrade

sudo apt install curl

cd
curl -OL https://github.com/microsoft/onnxruntime/releases/download/v1.17.1/onnxruntime-linux-x64-1.17.1.tgz

tar -zxvf onnxruntime-linux-x64-1.17.1.tgz
rm onnxruntime-linux-x64-1.17.1.tgz

3. Build onnx model

Add your username to --ort-path.

cd
cd ten-vad/examples_onnx
./build-and-deploy-linux.sh --ort-path /home/<username>/onnxruntime-linux-x64-1.17.1

4. Find onnx model

ls -halt build-linux/x64/onnx_model

5. Demo script

Use the python script from PR. Identify your WAV file in --input.

python3 vad_demo.py --model build-linux/x64/onnx_model/ten-vad.onnx --input ~/<your.wav>

6. Additional resources

• TEN VAD repo PR has ONNX model demo using Python extension module for the TEN VAD C/C++ library. Builds tested on Linux and macOS.

Next steps

• Download onnxruntime v1.22.0 version. Tested in TEN VAD repo PR.
• Download onnxruntime v1.22.0 for GPU version.