Jetson Nano and TX2 performance

Accelerated Computing CUDA CUDA Programming and Performance
1

Hello,

How would you rate Jetson Nano performance? We’re trying to detect people in video recordings. We use the YOLO Tiny algorithm, but its maximum on Nano is only 1 fps! We have also tried to run the Open Pose algorithm, but the it does not start at all.
The data about Jetsons performance provided by NVIDIA seem to be unreachable.
Do you have similar results? Is the nano performance so bad? Jetson TX2 isn’t much more efficient in our situation too.

What are yours experiences?

Best regards,
Mateusz

2

I think a question like this will have a more focused audience if you ask it on the Jetson forum(s).