Some questions about python deepstream_test3

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson)
**• DeepStream Version5.1
• JetPack Version (4.5)
**• TensorRT Version7.x

The pipeline structure of example 3 is as follows:

uridecodebin --> nvstreammux --> queue1 --> nvinfer(pgie) --> queue2 --> tiler --> queue3 --> nvvideoconvert --> queue4 --> nvdsosd --> queue5 --> nveglglessink

Question one:why so many queues are used in this pipeline?What are their functions?
Question two:In order to reduce the use of memory, I want to turn off the screen display. What should I do?

You can replace the nveglglessink with fakesink