@WongKinYiu @AlexeyAB
Thanks for this wonderful works.
I have trained an object detection model to predict road signs such as speed limits on custom dataset. Found that the predicted bounding box are smaller that the actual GT bounding box. I have used the p6 model and the hyp.scratch.1280.yaml config
@WongKinYiu @AlexeyAB Thanks for this wonderful works.
I have trained an object detection model to predict road signs such as speed limits on custom dataset. Found that the predicted bounding box are smaller that the actual GT bounding box. I have used the p6 model and the hyp.scratch.1280.yaml config
Any idea on what could be the possible reason?