Closed TheShadow29 closed 5 years ago
Hi @TheShadow29, the stanford_head
field is the index of the head token output by the Stanford dependency parser. It is the same as the head
field used in the standard CoNLL-U format files. If you are only using this repo, you can ignore this field, because the position-aware attention model does not need the dependency information.
I am trying to work with a different dataset. I can understand the other keys in the dictionary but I am clueless about
stanford head
. Could you suggest how to generate that?