About Result - Githubissues

tae898 / erc

The official implementation of "EmoBERTa: Speaker-Aware Emotion Recognition in Conversation with RoBERTa"

MIT License

88 stars 23 forks source link

About Result #21

Closed XinyeDu1204 closed 2 years ago

XinyeDu1204 commented 2 years ago

The first time I run the code, the result on the test dataset is more than 60. But when I ran the model again, the f1 score of the model was only about 8%, but I didn't modify any code. Do you know what caused this? Thank you!

tae898 commented 2 years ago

Hi,

Can you provide me with more information?

What's the data? MELD or IEMOCAP?
What code did you actually run?
Are there any codes, logs, messages that I can take a look at?

Thanks!

XinyeDu1204 commented 2 years ago

Thank you! I have solved this problem. Because I ran this project again in the result of the last run, which led to this problem.