Replace model in Deepstream SDK on Jetson

Autonomous Machines Jetson & Embedded Systems Jetson TX2
1

I asked this question but not able to open the question link so posting again
I want use YOLO darkflow with Deepstream on Jetson TX2.
Can I do it?
wayne zhu replied:-

Pharate,
For Yolo, our TRT can’t support all layer directly.
So it need IPlugin in TRT, which is not supported in our gst-nvinfer plugin.

So for this network, pls break down the whole pipeline now.
for example:
RTSP-> decode → VIC → APPSINK to get frame data, then use TRT + IPlugin to do inference.

Thanks
wayne zhu

Anybody have tried to do so?

2

Can somebody elaborate the flow a bit?
Thank you,
Komal

3

Hi pharatekomal,
Which video source are you using, camera or IP camera?
If IP camera, on tegra side, most probably you are using RTSP, then you can use following pipeline(with our gstreamer(deepstream solution)):
RTSP → H264parse → H264/5 decode → nvvidconvert → Appsink → Get raw frame,
Then in your APP, you can push raw frame to TRT.

If you using camera, it depends on which interface you are using, if V4l2 interface, you can get raw frame, if gstnvcamerasrc, then:
gstnvcamerasrc → appsink → get raw frame, Then in your APP, you can push raw frame to TRT.

If you have detail question, pls let me know.