I'm using coreNLP for coreference resolution from command line. I need the system to annotate singletons (single mentions) to be able to evaluate the output on my data, which has them annotated. I know that the system was trained on Ontonotes, which doesn't have singletons, but I found this parameter output.printSingletonEntities which is supposed, I guess, to add singletons annotation in some heuristic way.
output.printSingletonEntities : Boolean. Whether to print singleton entity mentions in coreference output.
However, adding it into the parameters file changed absolutely nothing. What am I doing wrong?
This is my parameter file:
annotators = tokenize,ssplit,pos,lemma,ner,parse,coref
tokenize.whitespace = true
tokenize.options = tokenizePerLine
ssplit.newlineIsSentenceBreak = always
coref.algorithm = neural
filelist = /Users/veronika/Dropbox/data/preprocessing/filelist.txt
output.printSingletonEntities = true
I also tried to call it from command line directly, but the annotation stays the same.
java -cp "" -Xmx10g edu.stanford.nlp.pipeline.StanfordCoreNLP -props /Users/veronika/Dropbox/data/sampleProps.properties -output.printSingletonEntities true*
Thank you in advance for your advices!
