Add documentation on how to use a model after it is trained.

[gilljon]

How do we actually use the trained model? How would we feed in text and get the graph associated with that text?

[f1amigo]

I have the same problem as well and would appreciate any help on this. Thanks!

[imelnyk]

I have updated main.py and README to include inference step. This is just a single input inference, which converts one input text into graph, using earlier trained model. All you have to do is to set --run inference and --inference_input_text "your input text".