I used the track function in mmdet, and I trained the detector and reid.
Initialize : trackor = init track model (
args.tracking config, args.det checkpoint, args.reid checkpoint, None, device = args.device )._
In inference mot_, I only printed the result of the detector, which is different from what I expected.
Next, I use the same detector weight and configuration file, where the configuration file is the same as the detector part model configuration in the track.
Initialization : detector = init detector (
args.det config, args.det checkpoint, device = args.device )_
The output of inference detector_ is different from the result of detector in the above track.
I used the track function in mmdet, and I trained the detector and reid. Initialize : trackor = init track model ( args.tracking config, args.det checkpoint, args.reid checkpoint, None, device = args.device )._ In inference mot_, I only printed the result of the detector, which is different from what I expected. Next, I use the same detector weight and configuration file, where the configuration file is the same as the detector part model configuration in the track. Initialization : detector = init detector ( args.det config, args.det checkpoint, device = args.device )_ The output of inference detector_ is different from the result of detector in the above track.