Migrating From JetPack 5 to JetPack 6

From RidgeRun Developer Wiki



Previous: Introduction/Beyond Compute Stack Index Next: Migrating from JetPack 5 to JetPack 6/Debian Package Method






JetPack is NVIDIA's comprehensive software development kit (SDK) designed to simplify the deployment of AI applications on embedded and edge devices. It provides a unified framework encompassing the essential tools, libraries, and technologies required for developing on NVIDIA's Jetson platform. JetPack offers support for a wide range of functionalities, including deep learning, computer vision, GPU computing, multimedia processing, and more, thereby enabling developers to harness the full potential of NVIDIA's powerful hardware.

Migrating from JetPack 5 to JetPack 6 involves transitioning to the latest version of this SDK, which promises enhanced features and software capabilities. The migration process is essential for developers looking to leverage the advancements in AI and edge computing provided by JetPack 6, such as updated versions of CUDA, cuDNN, TensorRT, and other critical libraries. This transition not only ensures compatibility with the latest NVIDIA hardware but also provides access to cutting-edge tools and optimizations that can significantly enhance the development and deployment of AI-powered applications. The following sections will help you navigate the migration process. It covers two primary scenarios: upgrading from JetPack 5 to JetPack 6, and updating minor versions within JetPack 6.

  • Debian Package Method section shows the high level steps on how to update the minor versions of JetPack 6 using the apt command.
  • Over-the-Air Update Method section shows high level steps on how to update the BSP on a Jetson device by writing pre-generated images to its partitions. This method allows to update from JetPack 5 to 6, as well as updating minor versions of JetPack 6.



Previous: Introduction/Beyond Compute Stack Index Next: Migrating from JetPack 5 to JetPack 6/Debian Package Method