My personal adventures with the turing pi rk1 with individual projects placed in submodules.
This repository serves as a high-level overview of the projects I am working on with the RK3588. Further information and notes about a project can be found in the README files of the corresponding submodules.
It is also my point of reference for anything found online that I might want to revisit later. The README files of the individual projects only contain directly relevant links to resources used in the project.
These are the individual projects that I think about doing with the RK3588. I will be updating this list as I go along.
Once a project has been started it will be checked off here and a link to the corresponding submodule will be added. A checked off item in this list does not mean it is complete, check out the submodule for more information.
The list is in no particular order, allthough I try to keep it somewhat organized/coherent.
- Setup a custom dietpi image for the RK3588
- Test out the RK3588 open-source NPU driver on a single node
- Setup automation for image building and deployment for the RK3588
- Getting the RK3588 open-source NPU driver working on all nodes
- Running a ML model (YOLOv8 or later) on RK3588 open-source driver (or at least getting it running)
- Setting up a cluster with the turing pi 2 (preferably using docker swarm CE)
- Running a distributed ML model on the RK3588 cluster
- Setup stable diffusion with web on the RK3588 cluster
- Setup a CI/CD pipeline for the RK3588 cluster
Someday there might be success visible here, but as of right now there isn't much to be seen.
The following is a list of references that I have found and collected. I try to keep it somewhat categorized to the topics that I am working on.
- Author's Blog on rockchip NPU kernel driver
- Review process of NPU kernel driver (Mailing list)
- Pull Request of NPU kernel driver
- RKNN-Toolkit2 (GitHub)
- airockchip - RKNN LLM (GitHub)
- Rockchip RK3588 performance with llama.cpp (GitHub issue)
...
...