Added encoding, recall total calculation and progress bar

NorskRegnesentral / text-anonymization-benchmark

Annotated corpus + evaluation metrics for text anonymisation

MIT License

50 stars 7 forks source link

Added encoding, recall total calculation and progress bar #1

Closed CrisesUrv closed 3 years ago

CrisesUrv commented 3 years ago

Added parameter encoding="utf-8" when open the gold standard json file to avoid UnicodeDecodeError Added the calculation and printing of the recall total Added a text bar to show the progress of the calculation of the precision (useful for bert weighting because it takes quite a long time)