I’m getting multiple bounding box detections over image.Converted my own trained model on my PC and used this converted model on VIM4
Ubuntu 18.04.6 LTS
Python 3.8.0
ultralytics 8.3.155
on my VIM4 board its:
Python 3.10.12
Ubuntu 22.04.5 LTS
@176gauravsharma have you verified if the default example works ?
yes, its working fine
Hello @176gauravsharma ,
Have you verified your model output is right like doc? Have you modified the LISTSIZE and NUM_CLS to your model in demo codes?
If they are all right, please provide your ONNX model and adla model. I will help you solve the problem.
My model has 6 classes & make changes for forward and forward_export function in head.py
onnx converted model output on netron app is:
Hello @176gauravsharma ,
Your model output is right and could you provide your ONNX model and adla model?
@Louis-Cheng-Liu Thanks for your support actually have to change output_shape parameter in yolov8.nn_inference according to my model. now i’m getting correct inference on VIM4 NPU.
