- This project aimed at evaluating the effects of making short edits on a text, like news headlines and, does this short modification make the text more humourous, i.e. from non-funny to funny.
- Project made using standard BERT transformer, made available through HuggingFace ibrary.
- This was a regression task and an RMSE of 0.57 was obtained.
- It is slightly above the base line score.
- A lot more samples and restructuring of the data is required to obtain a low RMSE score, as this task is quite tricky.
-
Notifications
You must be signed in to change notification settings - Fork 0
Using Transformer for Regression tasks (scoring) in NLP
License
HargovindArora/Assessing_the_Funniness_of_Edited_News_Headlines
Folders and files
Name | Name | Last commit message | Last commit date | |
---|---|---|---|---|
Repository files navigation
About
Using Transformer for Regression tasks (scoring) in NLP
Resources
License
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published