Could you please update the detailed README to match the experiment in the paper

[Zeng1998]

I want to apply a new dataset in "only sherlock" and the fully "sato" model.

I found train_sherlock.py and train_CRF_LC.py

I run train_sherlock.py and --topic=None. Is the result of only sherlock?

I run train_sherlock.py with topic and train_CRF_LC.py with topic. Is the result of the fully sato model?

And, what does train_sherlock.py with topic mean?

Thanks