bert-finetuned-ner

Dataset Used

This model was trained on the CoNLL 2003 dataset for Named Entity Recognition (NER) tasks.

The dataset includes the following labels:

• O, B-PER, I-PER, B-ORG, I-ORG, B-LOC, I-LOC, B-MISC, I-MISC

For detailed descriptions of these labels, please refer to the dataset card.

Model Training Details

Training Arguments

• Model Architecture: bert-base-cased for token classification
• Learning Rate: 2e-5
• Number of Epochs: 20
• Weight Decay: 0.01
• Evaluation Strategy: epoch
• Save Strategy: epoch

Additional default parameters from the Hugging Face Transformers library were used.

Evaluation Results

Validation Set Performance

• Overall Metrics:
  • Precision: 94.44%
  • Recall: 95.74%
  • F1 Score: 95.09%
  • Accuracy: 98.73%

Per-Label Performance

Test Set Performance

• Overall Metrics:
  • Precision: 89.90%
  • Recall: 91.91%
  • F1 Score: 90.89%
  • Accuracy: 97.27%

Per-Label Performance

How to Use the Model

You can load the model directly from the Hugging Face Model Hub:

from transformers import pipeline

# Replace with your specific model checkpoint
model_checkpoint = "Prikshit7766/bert-finetuned-ner"
token_classifier = pipeline(
    "token-classification", 
    model=model_checkpoint, 
    aggregation_strategy="simple"
)

# Example usage
result = token_classifier("My name is Sylvain and I work at Hugging Face in Brooklyn.")
print(result)

Example Output

[
   {
      "entity_group":"PER",
      "score":0.9999881,
      "word":"Sylvain",
      "start":11,
      "end":18
   },
   {
      "entity_group":"ORG",
      "score":0.99961376,
      "word":"Hugging Face",
      "start":33,
      "end":45
   },
   {
      "entity_group":"LOC",
      "score":0.99989843,
      "word":"Brooklyn",
      "start":49,
      "end":57
   }
]