Ner Bert Base Cased Ontonotesv5 Englishv4

🚀 (NER) bert-base-cased : conll2012_ontonotesv5-english-v4

This is a Named Entity Recognition (NER) model based on bert-base-cased, which has been fine - tuned on the conll2012_ontonotesv5 dataset in the english-v4 version. It can effectively identify named entities in text. For more information, check out the NER-System Repository.

📚 Documentation

Tags

• token-classification

Task IDs

• named-entity-recognition

Datasets

• djagatiya/ner-ontonotes-v5-eng-v4

Widget Example

• text: "On September 1st George won 1 dollar while watching Game of Thrones."

📊 Evaluation

The model has been evaluated with the following metrics:

• Precision: 87.85
• Recall: 89.63
• F1-Score: 88.73

⚠️ Important Note

Check out this eval.log file for detailed evaluation metrics and classification report.

               precision    recall  f1-score   support

    CARDINAL       0.86      0.87      0.86       935
        DATE       0.84      0.88      0.86      1602
       EVENT       0.65      0.67      0.66        63
         FAC       0.69      0.71      0.70       135
         GPE       0.97      0.93      0.95      2240
    LANGUAGE       0.76      0.73      0.74        22
         LAW       0.54      0.55      0.54        40
         LOC       0.73      0.80      0.76       179
       MONEY       0.87      0.90      0.88       314
        NORP       0.93      0.96      0.94       841
     ORDINAL       0.80      0.87      0.83       195
         ORG       0.88      0.90      0.89      1795
     PERCENT       0.88      0.90      0.89       349
      PERSON       0.94      0.95      0.94      1988
     PRODUCT       0.62      0.76      0.69        76
    QUANTITY       0.74      0.81      0.77       105
        TIME       0.61      0.67      0.64       212
 WORK_OF_ART       0.56      0.66      0.61       166

   micro avg       0.88      0.90      0.89     11257
   macro avg       0.77      0.81      0.79     11257
weighted avg       0.88      0.90      0.89     11257