training data used for coreference

[Matthew Gerber]

Hello,

Does anyone know what data was manually annotated and used to train the named entity and coreference models? I'm doing experiments over particular sections of the Penn TreeBank (23 and 24), and would like to avoid using tools that train over these sections, for obvious reasons.

Thanks!
Matt Gerber

[Matthew Gerber]

*bump*

Is anyone familiar with the issue I raised?

[Thomas Morton]

Hi,
   It's trained on Penn Treebank sections 00 and 01.  Hope this helps…Tom

[Matthew Gerber]

Thanks, Tom! It does help.