Improve implementation of jetson inference

I apologize if everything is very basic. I am new with a Jetson Nano, I just received it.

  1. I have a Jetson Nana and my camera is 2K. That is, when the camera stream is displayed, resources are consumed. Is there a way to make it not show?

  2. I would like to know if I can make the dimensions of the camera smaller so that the reading is faster.

  3. Is it possible to see the code that is inside the docker? That is, I not only want to use it and now, I want to have control of it.

Greetings.

Update:

Solution 2: jetson-inference/aux-streaming.md at master · dusty-nv/jetson-inference · GitHub
–input-width=1920 --input-height=1080 csi://0

Hi,
We would suggest try gstreamer or jetson_multimedia_api. You may refer to the document:
gstreamer user guide
Jetson Linux API Reference: Main Page

You may try this command and check if you can see camera preview:

$ gst-launch-1.0 nvarguscamerasrc ! nvoverlaysink

And you can select the sensor mode with low resolution, or capture data in high resolution and then downscale to low resoluton through nvvidconv plugin.