jetson-containers

jetson-containers is a Docker container build system for developers running local AI on NVIDIA Jetson. It assembles GPU software and its dependencies into containers for workloads that need a compatible CUDA stack.

• AI frameworks: Build containers with packages including PyTorch, TensorFlow, JAX and ONNX Runtime, alongside tools such as JupyterLab and DeepStream.
• GPU library compatibility: Rebuild the stack around a different CUDA version or select compatible versions of PyTorch, TensorRT and cuDNN when a project requires a particular combination.
• Reusable build downloads: A package cache reduces repeated downloads and compilation, with an optional local PyPI and APT caching server.

The containers run on NVIDIA Jetson hardware and include Ubuntu images with CUDA support. ARM SBSA systems such as NVIDIA GH200 and GB200 are also supported. Container builds can retrieve images and dependencies from external servers. Compatibility depends on the device's JetPack release and CUDA driver; Jetson Orin's ARM SBSA CUDA path has additional JetPack and driver requirements.