Hello, I need help on modifying the output of the model.
I’m using YOLOX as the model, and it has the ff. output: labels, dets.
It should be straight forward on how to interpret the output but I am constrained on a embedded device that only accepts these kind of output: locations, classes, scores, num_of_dets kind of similar to SSD MobileNetv1.
I also tried doing the SSD model but the accuracy is low compared to the YOLOX trained with same dataset.