custstu/yolov3-tiny_3l-TensorRt-for-tableware-data

using darknet yolov3-tiny_3l model to detect tableware

yolov3-tiny_3l TensorRt for tableware data

Using darknet yolov3-tiny_3l model to detect tableware on nvidia jetson nano. Add a function "_process_feats_and_filter_boxes" to optimize time consumption, and the inference time of one image is about 60 milli seconds on jetson nano.