thangnm99

I think it's because the trained model is trained on a larger scale image. You can see the input size in the config file of the detection model. I do...