Sizing Storage for GPU Jetson Xavier NX

Hello guys,

Thank you for Welcoming me to be part of this community.
Currently i have challange to sizing the capacity storage on video analytic, because i have deploy it on-premise so i need to precisely the requirement for storage.

For detail environment i want to analyze is about people movement, deck movement and counting how many boxes already moved for 1 side to another side.

Beside that, i need to store the output for inference result (categorized) to playback for showing the past activity.

If there anyone have same case, please share your idea.
Thank you very much 😉

Hi,
Have you tried deepstream-app? For detection and classification, you may try to run

/opt/nvidia/deepstream/deepstream-5.0/samples/configs/deepstream-app$ deepstream-app -c source4_1080p_dec_infer-resnet_tracker_sgie_tiled_display_int8.txt

The default models are in

/opt/nvidia/deepstream/deepstream-5.0/samples/models

Yo may try the default deepstream-app as reference and apply your models.