Install XenseSDK

1. Create a Python environment

Anaconda with Python 3.9 is recommended. Current examples use 3.9.19; Python 3.10 is also a compatibility option.

conda create -n xenseenv python=3.9.19
conda activate xenseenv

2. Install CUDA and cuDNN

GPU inference requires CUDA Toolkit and cuDNN versions compatible with onnxruntime-gpu. The legacy documentation validated CUDA 11.8 with cuDNN 8.9.2.26; for newer environments, follow the release-package notes and GPU-driver compatibility requirements.

conda search cudnn
conda search cudatoolkit
conda install cudnn==8.9.2.26 cudatoolkit==11.8.0

On Linux, if the runtime cannot find shared libraries:

export LD_LIBRARY_PATH="$CONDA_PREFIX/lib:$CONDA_PREFIX/lib64:$LD_LIBRARY_PATH"

Without an NVIDIA GPU, you can first create the sensor with use_gpu=False to verify communication and the data path using CPU inference.

3. Install the package

pip install xensesdk

For a local wheel, select the file matching Python, OS, and architecture:

pip install xensesdk-x.y.z-cp39-cp39-win_amd64.whl

When available in the release package, install the visualization or full dependency set with:

pip install 'xensesdk[viz]'
pip install 'xensesdk[full]'

Verify after installation

Run python -c "from xensesdk import Sensor; print(Sensor.scanSerialNumber())". If the result is empty, check USB connectivity, device permissions, compute-unit networking, and sensor configuration files.