You’ve got a ton of good documentation for working with custom models, but (due to my own middling coding abilities) I’ve struggled to convert models from onnx to hef using a VM or Colab notebook.
Any chance the Hailo crew (or maybe kind-hearted community members?) could do a couple walkthroughs of very common workflows? For example, taking a YOLOv11n model that’s been finetuned on 3 or 4 classes and then exporting, generating a .har file, quantizing, and compiling to .hef.
Could be limited office hours with some sort of signup in advance, or maybe a YT Live where you field some questions.
Something to consider. Tx for building such an exciting product for us AI hobbyists!
Thank you for bringing this up! You can follow our YOLOv5 guide as a reference since the process is quite similar for YOLOv8 or YOLOv11. The core steps and principles remain largely the same between these versions.
That said, I’ve shared your feedback with our team about creating specific documentation for YOLOv11. On a related note - stay tuned, as we have an exciting announcement coming up for hobbyists regarding the YOLO family of models! We’ll share more details once everything is finalized.