I'm using object_detection_demo_yolov3_async from OV 2020.2. I downloaded input video from here. The models were downloaded using openvino\inference_engine\demos\python_demos\object_detection_demo_yolov3_async\models.lst
The demo runs but output has bounding boxes that are not proper. Please see attached. Similar results with other inputs. The second attached file (Capture2.jpg) is upper-left snip of first attached file.
thanks for letting us know. Could you please also specify what particular model (yolo v1, v2 tiny or yolo v3) does not work as expected? It also would be useful to provide picture used as input for the demo.
Can't reproduce the issue on specified video using OpenVINO 2020.2 on Windows and yolo-v3-tf model from Open Model Zoo and python object_detection_demo_yolov3_async. Could you please provide more details on how do you run demo, what is OS and device you run inference on?