MrRobson9/gpt2-ner-conll2003-english

GPT-2 Fine-Tuned on CoNLL2003 for English Named Entity Recognition (NER)

This model is a fine-tuned version of GPT-2 on the CoNLL2003 dataset for Named Entity Recognition (NER) in English. The CoNLL2003 dataset contains four types of named entities: Person (PER), Location (LOC), Organization (ORG), and Miscellaneous (MISC).

Model Details

• Model Architecture: GPT-2 (Generative Pre-trained Transformer)
• Pre-trained Base Model: gpt2
• Dataset: CoNLL2003 (NER task)
• Languages: English
• Fine-tuned for: Named Entity Recognition (NER)
• Entities recognized:
• PER: Person
• LOC: Location
• ORG: Organization
• MISC: Miscellaneous entities

Use Cases

This model is ideal for tasks that require identifying and classifying named entities within English text, such as:

• Information extraction from unstructured text
• Content classification and tagging
• Automated text summarization
• Question answering systems with a focus on entity recognition

How to Use

To use this model in your code, you can load it via Hugging Face’s Transformers library:

from transformers import AutoTokenizer, AutoModelForTokenClassification
from transformers import pipeline

tokenizer = AutoTokenizer.from_pretrained("MrRobson9/gpt2-ner-conll2003-english")
model = AutoModelForTokenClassification.from_pretrained("MrRobson9/gpt2-ner-conll2003-english")

nlp_ner = pipeline("ner", model=model, tokenizer=tokenizer)
result = nlp_ner("John lives in New York and works for the United Nations.")
print(result)

Performance

accuracy	precision	recall	f1-score
0.973	0.783	0.840	0.810

License

This model is licensed under the same terms as the GPT-2 model and the CoNLL2003 dataset. Please ensure compliance with all respective licenses when using this model.