Duplicated spark DAG graph by beam

Hi, 
We are using the beam to read the stream from Kafka and using spark as the runner, and I found our spark graph looks very strange.
 
For each batch stream, it will generate 3 stages, one of them of our actual work, that I can understand.
 
Another two is kind of duplicated, you can take a look at the attached graphs:

 
 

I don't see any reason we display two graphs but one of them include another one.
 
Br,
Tim