docs: Add high-performance C++ Edge AI port to README - #443
Conversation
|
|
@MatinRafiei mind check this pl ^^ 🦝 |
Hi @Borda, I have just reviewed and signed the CLA. Let me know if there is anything else you need from me for this PR! Best, |
|
@SkalskiP pls have a look ^^ 🦝 |
There was a problem hiding this comment.
Pull request overview
Adds a new README section to highlight a community-maintained C++ port of the trackers library for edge/embedded deployments, helping users discover non-Python options when targeting devices like Raspberry Pi/Jetson.
Changes:
- Add a “Community Ports” section to the README.
- Include a bullet linking to an external C++ edge-oriented port repository.
| ## Community Ports | ||
|
|
||
| - **C++ Edge Port:** A highly-optimized, hardware-agnostic C++ translation of this repository (BoTSORT, ByteTrack, OC-SORT and SORT) designed for real-time Edge AI deployments (Raspberry Pi, Jetson) is available at [EdgeTrackers-CPP](https://github.com/MatinRafiei/roboflow-trackers-cpp.git). | ||
|
|
|
"@Borda Hi Borda, following up on the status of this pull request. Please let me know if there are any updates or if further changes are required on my end." |
Hi Roboflow Team,
First, thank you for building and maintaining such a fantastic, clean tracking framework!
I recently completed a project to entirely port the core mathematical tracking architecture (BoTSORT, ByteTrack, OC-SORT, and SORT) into a highly-optimized, standalone C++ library designed specifically for Edge AI hardware (Raspberry Pi, NVIDIA Jetson, etc.) where Python overhead can be a bottleneck.
The library maintains mathematical parity with your implementations, utilizes
cv::Matfor seamless OpenCV/Edge NPU integration, and is hardware-agnostic.Since many developers deploying Roboflow models to the Edge eventually need to transition from Python to native C++, I thought it would be highly beneficial to add a link to this community port in the README.
Thank you for your time and the excellent foundational work!