Please provide complete information as applicable to your setup.
• Hardware Platform (Jetson / GPU): GPU
• DeepStream Version: 5.0
The way in which the Python APIs are currently designed (specific reference) is very C-esque in nature right now.
Would be very convenient if, for instance, in the above reference, the entire while loop is replaced with a single pyds.NvDsObjectMetaList.cast(frame_meta.obj_meta_list) that returns a numpy array with relevant numeric data from here rather than running pyds.NvDsObjectMeta.cast(l_obj.data)within a while-loop. Running it within a while loop in Python also slows down the Python application as a whole, which can be blocking for the Gstreamer pipeline