The weighted ROUGE scores should been calculated by two parses, but in the function generate_tgt_parse() the code is "possible_drawn = step2_rouge(level_, freq, src_lines[i], level)[3]". Is the src_lines[i] parameter an error and should be replaced with src_pure_parses? Are there any errors in the results obtained in the paper?
Looking forward to your reply!
The weighted ROUGE scores should been calculated by two parses, but in the function generate_tgt_parse() the code is "possible_drawn = step2_rouge(level_, freq, src_lines[i], level)[3]". Is the src_lines[i] parameter an error and should be replaced with src_pure_parses? Are there any errors in the results obtained in the paper?
Looking forward to your reply!