how can i use trt-yolo-app to run ourself yolo network with trained weights file?

now,my problem is : use nvidia offical yolo models run ok, but when use yolo network,then trained ourself images,then trt-yolo-app infer nothing!!, what is the problem,help me!!


Which YOLO model do you use? Ex. v2, v2_tiny, v3, v3_tiny?
Could you check if your .cfg file is as same as the architecture downloaded from this script?