Closed HaishuoFang closed 3 years ago
I printed the precision and recall of all relations, and the result is same as the official script.
Does the row mean predicted labels or true labels? I think it's a bit weird
I review my code and I think that row means the predicted labels, maybe the Perl script is wrong, but it does not affect the final F1.
If rows of confusion matrix are prediction and columns are labels, it seems they are misplaced.
https://github.com/onehaitao/Att-BLSTM-relation-extraction/blob/90db24659ec92442c0693cd086af8ce34691dc50/evaluate.py#L31