How to write a parallel prediction when using both model in deepbinner classify ? 

Dear @rrwick 
Compare to use start model only and both model , it seems to spend double time for model prediction.  
For my research, I need to do it parallel, because I have 20 models. 
Can you give me some suggestion, if I try to modify classify.py script?
My GPU card  are RTX 2080 and GTX1080.
Thanks    
