I used pretrained SpanBERT_base model to evaluate on gap-test.tsv after converting to jsonlines and got very low F1 (Average F1 : 17.53% , Average precision (py): 10.64% Average recall (py): 56.61% ). So, wanted to try on OntoNotes Dataset. Is it possible to get a sample of that ?
I used pretrained SpanBERT_base model to evaluate on gap-test.tsv after converting to jsonlines and got very low F1 (Average F1 : 17.53% , Average precision (py): 10.64% Average recall (py): 56.61% ). So, wanted to try on OntoNotes Dataset. Is it possible to get a sample of that ?