The given notebook example produces alignments/parses that are invalid for gcm process #16

Dear maintainers,
I ran the example notebook in this repository and while the notebook has provided that corpus that seems to produce good alignments/parses for the gcm stage, it is not the case when i ran the notebook. So I presume it is highly likely that there is some dependency issue. Can you please look on this problem? Thanks