training a New Spanish model from scratch

Hi there
I am trying to create a new Spanish large model from scratch, so far I manage to import a large FastText and POS tagging training however when i try to convert the wikiNER dataset (https://figshare.com/articles/Learning_multilingual_named_entity_recognition_from_Wikipedia/5462500) I keep getting this:

"The iob/iob2 file is not formatted correctly. Try checking whitespace and delimiters."

ValueError: The iob/iob2 file is not formatted correctly. Try checking whitespace and delimiters.

could you please help me with that issue?
thanks
Angelo