Hi Nvidia Team,
I want to do Deep Learning Inference on Jetson Nano with 8 Cameras(Object Detection Model with Mobilent SSD). Does it handle 8 camera streams? What are the ways to make it work?
Does Deepstream is only an Option? Or are there any other ways to achieve this?
And what are the factors that affect Multistream Inference on Edge Devices with any Computer Vision based Models?
Thanks,
Darshan
I am looking forward to your reply.
Thanks
Hi,
We would suggest use DeepStream SDK. You can try with default sample config:
nvidia@nvidia-desktop:/opt/nvidia/deepstream/deepstream-5.1/samples/configs/deepstream-app$ deepstream-app -c source8_1080p_dec_infer-resnet_tracker_tiled_display_fp16_nano.txt
Modify to 8 RTSP sources per:
DeepStream Reference Application - deepstream-app — DeepStream 6.1.1 Release documentation
And replace resnet10 with your model.