Skip to content

Conditional Random Field (CRF) model for sequence labeling as the baseline model for Quote Detection task.

License

Notifications You must be signed in to change notification settings

Darg-Iztech/quote-detection-crf

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Quote Detection CRF

Conditional Random Field (CRF) model for sequence labeling as the baseline model for Quote Detection task.

Run CRF baseline:

Example 1:

python qtask_baseline.py --dataset='T50' --iter=500 --from_scratch

Example 2:

python qtask_baseline.py --dataset='MOVIE' --iter=500

Calculate ROUGE scores manually:

pip install rouge

rouge -f data/pred_quotes.txt data/true_quotes.txt --avg

Results for 500 iterations and 5-fold cross-validation:

Metric T50 MOV
R1 20.28 ± 2.99 26.42 ± 0.13
R2 14.38 ± 2.88 20.72 ± 0.32
RL 19.38 ± 3.03 25.75 ± 0.20

References:

https://github.com/arielsho/SemEval-2020-Task-5

https://github.com/pltrdy/rouge

About

Conditional Random Field (CRF) model for sequence labeling as the baseline model for Quote Detection task.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published