Skip to content

Latest commit

 

History

History
3 lines (3 loc) · 725 Bytes

bert_classifier.md

File metadata and controls

3 lines (3 loc) · 725 Bytes

In order to use the BERT classifier, first install gdown using pip install gdown, then proceed as below considering the model you want to use:

  1. To use "distilbert_base_uncased", first download the trained model using gdown 132xBqgzsccaZg-E5hmsihoTCvJI7jQUD and unzip the file, then set the variable MODEL_TO_USE to "distilbert_base_uncased" in bert_classify.py and run it. This model achieves an F-score of 0.09 and an accuracy of 40%
  2. To use "bert_base_uncased", first download the trained model using gdown 154XTPXWaHnALT3IX8tX32J7-ht_Cz3vI and unzip the file, then set the variable MODEL_TO_USE to "bert_base_uncased" in bert_classify.py and run it. This model achieves an F-score of 0.13 and an accuracy of 31%