Openpilot 0.10.1 introduces the North Nevada Model, featuring major improvements to the World Model architecture. The system now infers 6 degree of freedom ego localization directly from images, removing the need for external localization inputs. This reduces over-constrained data and opens the door for future self-generated imagery.
To support this change, the autoencoder Compressor was upgraded with masked image modeling, switched from CNN to Vision Transformer architecture, and the World Model itself was scaled from 500 million to 1 billion parameters. All models now train on a much larger dataset of 2.5 million segments, up from 437,000, covering more vehicles, countries, and driving scenarios.
The UI has been completely rewritten, moving from Qt/Weston to Python with raylib. This reduces code complexity by about 10,000 lines, cuts boot time by 4 seconds, lowers GPU usage, and simplifies development.
Finally, the Driver Monitoring Model’s training infrastructure has been streamlined with dynamic data streaming, though the model’s functionality remains unchanged.
Installation is pretty straight forward, usually it looks like plugging into your OBD2 port, running a cable up to your window, and attaching an adaptor into your camera system. It basically sets up like a dash cam, but with more integration into the vehicle.
The software and device currently supports 325+ cars. I think the car needs to have electronic steering and the automatic distance keeping features. Unfortunately it doesn’t look like Volvo is supported, but the project is open source and can always grow. There are also community forks, I am currently using SunnyPilot.
You can find out more on their website, https://comma.ai/ .
Thanks for the info!