Open shyMlemon opened 1 year ago
Sorry for close the wrong issue. I open it again. You can adapt the data file to fit this for loop. https://github.com/laiviet/ed-gated-gcn/blob/d6a9a9e305a034d6c2532503b68aee18fa6693d5/data_utils.py#L637
In which each sample representing a sentence, each with <sample id, list of tokens, list of labels, list of targets> list of tokens, list of labels, and list of targets are all word-level
target is just the index of the corresponding label (which will be used in cross entropy function)
what is the format of sample in ACE2005 datasets?