Open so-hyeun opened 3 years ago
Hello, thanks for doing interesting research.
The picture above is the result of reproducing the MELD dataset performance through UniF-BiAGRU. Of the results, is 57.77 the value that should be compared with the paper's F1 score?
Yes, 57.77 corresponds to the F1 score. You can see from the ``EmoTrain.py''.
Hello, thanks for doing interesting research.
The picture above is the result of reproducing the MELD dataset performance through UniF-BiAGRU. Of the results, is 57.77 the value that should be compared with the paper's F1 score?