I am working with a Raspberry Pi 5 that has no internet connectivity at all. It took me quite some time, but I successfully installed all the dependencies required by hailo-all.
However, when I try to run the examples from the hailo-rpi5-examples repository, I notice that they rely heavily on internet access—for example, installing Python packages via pip or cloning repositories from Git.
Is there a guide that lists all the dependencies needed to run Hailo on a fully offline Raspberry Pi? Alternatively, is there a small example project for object detection that can be run entirely offline? A step-by-step guide would be very helpful.
The examples from the Hailo repositories assume an environment with at least temporary internet access. That’s mainly because they need to pull dependencies, models, and supporting code during setup.
Running completely offline is possible, but typically only after you’ve completed development and installation on a connected system. I would recommend you.
Set up and test the environment on a system with internet access.
Record all dependencies.
Use tools such as pip download or local package mirrors to collect everything needed.
Transfer those files to the offline Raspberry Pi and install them locally.
All of our examples work offline. We do have different repositories to support different requirements and skill levels.
You don’t necessarily need Python at all. You can develop your application in C++. The following repo is more intended as a starting point to develop your own application.