This project aim to realize and optimize the yolo-tiny network on embedded platforms. These platforms including FPGA(Zynq), embedded GPU(Nvidia Jetson)
The project only has the inference functionality of YOLO-Tiny, the training functionality is not reserved, since it's not likly you want to train your network on these platforms.
The project derived from the Darknet framework, credit goes there.
To test on an image: ./darknet yolo test cfg/yolo-tiny.cfg yolo-tiny.weights IMAGEPATH
To test on video: .darknet yolo demo cfg/yolo-tiny.cfg yolo-tiny.weights VIDEOPATH
To test on webcame: .darknet yolo demo cfg/yolo-tiny.cfg yolo-tiny.weight
To visualize the net: ./darknet visualize cfgfile weightfile