Hey guys,
we would like to run our yolov8-pose models on the OAK-D Short Range PoE with ToF.
How would I convert the model into a functioning blob file?
Thanks!
Converting yolov8-pose to run on OAK-SoM-Pro
- Edited
FredWerner
Convert it like you would any other model: https://docs.luxonis.com/software/ai-inference/conversion/. Keep in mind to use NeuralNetwork node, since YoloDetection can only do detection.
Then you need to extract the output and provide a custom decoding function on host side.
Thanks,
Jaka
Additional note on this is that we are also working on adding support to FW directly, and will share more details on that in the following weeks.