[error] Mask is not defined for output layer with width '8'

Iqbalalwi

I tried to convert YOLOV5 training results to Blob and it worked. But when I apply to device-decoding (https://github.com/luxonis/depthai-experiments/tree/master/gen2-yolo/device-decoding) an error occurs

"[14442C10C1FA3BD400] [1.1] [973.371] [DetectionNetwork (1)] [error] Mask is not defined for output layer with width '8'. Define at pipeline build time using: 'setAnchorMasks' for 'side8'."
I am using 3 classes, input image 416 and I have adjusted it with JSON file.

can anyone help me solve this problem?

erik

Hi Iqbalalwi ,
DId you use tools.luxonis.com to get your .json, as it's mentioned in the Yolov5 training notebook?
Thanks, Erik

Iqbalalwi

it's saved me a lot!, tks