Khadas VIM3 custom one-class YOLOv3 inference issue

@Akkisony Can you show you cfg file ?

@Frank
https://drive.google.com/drive/folders/14wDTO81KBYYfyz0Jj0aWsCXWFixbHpwS?usp=sharing

please find the cfg file.

@Akkisony

#batch=1
#subdivisions=1
# Training
batch=64
subdivisions=16

you should use bahtch=1 subdivisions=1

@Frank Thanks. May be I forgot to change the parameter of cfg file while converting using tengine sdk. I will try and update you.
Thanks again!

@Akkisony OK, if it not work , I will try you cfg and weights

@Frank I was not successful, i still have the same problem when I try to do inference on NPU (video).

I changed batch and subdivision to 1, while converting the model to tmfile using tengine sdk.
However, during training it was 16 and 64.

https://drive.google.com/drive/folders/14wDTO81KBYYfyz0Jj0aWsCXWFixbHpwS?usp=sharing

Please find the model and cfg.

@Akkisony I will try it next week

@Frank Okay, thank you.
May be if you can just try running (through video) the converted model in the above link, you might get to know my problem I was talking about.

@Frank Did you try the model?

@Akkisony In my plan this week, but it has not been tested yet

@Frank Thanks. Can you pls make time and try it by today or tmrw?

@Akkisony Sorry, I have something urgent to deal with, I will try my best to try your problem tomorrow

@Frank thanks. Looking forward to your solution.

@Akkisony What is your model used to identify classification?

@Frank I am sorry, I dint get your question.

@Akkisony What is the class lists of your model ?

@Akkisony I use you model and using the code of my hand demo, it runs very well, I think your code processing has a problem.

@Frank Thanks. I am also using the demo code of tengine sdk, but I hv problem with Non-Max Supression. Can you please hv a look at the .cpp file attached in the link?
https://drive.google.com/drive/folders/14wDTO81KBYYfyz0Jj0aWsCXWFixbHpwS?usp=sharing

@Akkisony

@Frank