Nvbuf_utils: dmabuf_fd 1339 mapped entry NOT found

Hi all,
I am developing a Pedestrian detection application on Jetson Xavier NX which is supposed to process video coming from an Hikvision IP camera stream using rtsp.
In my pipeline, I constantly get a the following errors before the program crashes after maximum of 3 minutes:

nvbuf_utils: dmabuf_fd 1335 mapped entry NOT found
nvbuf_utils: NvReleaseFd Failed… Exiting…

What I have done so far:
(1) I have read through similar posts in the forum but none seem to be able to particularly solve my problem.
The strange part is that, the same exactly code runs on my Jetson Nano 4gb without showing the same error.
(2) I have installed different Jetpack versions on different SD cards to test the application.
(3) I have try out different jetson-inference branches to test the the application including: dusty-nv/jetson-inference “nvmm-disable” branch. and the error still pops up regularly and crashes the application after 3 minutes.

I am using pednet model for the pedestrian detection and I religiously follow all the instruction when I am building the jetson inference project from Source.

The logic of my code is presented below:

(1) InputURL = “hikvision camera rtsp url”
(2) input = jetson.utils.videoSource(inputURL, argv=[’–input-rtsp-latency=0’, ‘–input-codec=h264’, ‘–input-width=1280’, ‘–input-height=720’])
(3) models = ‘pednet’
(4) net = jetson.inference.detectNet(models, threshold=0.5)
while loop:
(5) video_stream = input.Capture() # video stream capture
(6) detections = net.Detect(video_stream, video_stream.width, video_stream.height)
(7) video_stream = jetson.utils.cudaToNumpy(video_stream) # so that opencv can display
(8) for pedestrian in detections: ## for loop to check for detection details
(9) cv2.imshow(‘Detection’, video_stream) # opencv imshow.

Please i need help to solve the error. Thank you