making the tracker faster

[Elad-Shmueli]

i was hoping someone could help me push faster processing from the trackers with the cost of lower accuracy.
i am not sure about the difference in models and configs and backbones but alex model runs fastest on my machine(jetson xavier NX),
it runs on about 30~fps, but i want more, and i dont mind losing performence.
can anyone please tell me what i can do to make it run faster, do i need to train the tracker on a small data set? or make the program use more GPU power?
any insight would be highly appreciated