Yolov5 on OAK-D Pro W PoE

StavanRupareliya

Hello,
I have Luxonis OAK-D Pro W PoE camera. I would like to run inference on OAK-D Pro W PoE using yolov5.
Is it possible?
Can you guide me How it is possible?

erik

Hi StavanRupareliya ,
Yep it's possible, and it's one of the most popular AI models that's used with our devices. I'd suggest checking yolov5 notebook and skipping training (if you already have the model), then proceeding to model conversion and deployment to the OAK camera. Thoughts?
Thanks, Erik

StavanRupareliya

Hello erik ,
I am trying to run yolov5 on OAK Camera got this error.

error -
[184430106187C51200] [172.25.3.57] [48.335] [DetectionNetwork(1)] [error] Mask is not defined for output layer with width '6'. Define at pipeline build time using: 'setAnchorMasks' for 'side6'.

erik

Hi StavanRupareliya , have you followed notebook linked above step-by-step?
Thanks, Erik

StavanRupareliya

Hi erik ,
Where i need to find yolov5s model JSON file ?

StavanRupareliya

Hi erik
Please guide me regrading Custom model Jeson file ..
ERROR-
[184430106187C51200] [172.25.3.57] [11.359] [DetectionNetwork(1)] [error] Mask is not defined for output layer with width '6'. Define at pipeline build time using: 'setAnchorMasks' for 'side6'.

erik

Hi StavanRupareliya ,
You download it from tools.luxonis.com - everything is written in the notebook I mentioned above. Please follow it.
Thanks, Erik

StavanRupareliya

Hello erik
it's not working for me.
Can you send me working JSON file for custom yolov5 model ...
ERROR - [DetectionNetwork(1)] [error] Mask is not defined for output layer with width '6'. Define at pipeline build time using: 'setAnchorMasks' for 'side6'

erik

StavanRupareliya , Please provide specifically step-by-step you have taken.

StavanRupareliya

Hello,

my error is -
error -
[184430106187C51200] [172.25.3.57] [48.335] [DetectionNetwork(1)] [error] Mask is not defined for output layer with width '6'. Define at pipeline build time using: 'setAnchorMasks' for 'side6'.

There is problem in JSON file.
Solution is here.
For this using Automatic Yolo export for OAKs tool for generating custom yolo model JSON file.
So, Eorror is solved.
Custom yolov5 model successfully ported in OAK camera.
Tool link - https://tools.luxonis.com/