BERT Named Entity Recognition with CoNLL-2003 Dataset

Named Entity Recognition (NER) is a crucial task in Natural Language Processing (NLP) that involves identifying and classifying named entities (e.g., persons, organizations, locations) within text data. This project demonstrates how to fine-tune the pre-trained BERT base uncased model for the NER task using the CoNLL-2003 dataset. The CoNLL-2003 dataset is a widely used benchmark dataset for the NER task, consisting of news articles from the Reuters Corpus. The dataset is annotated with four entity types: Person (PER), Organization (ORG), Location (LOC), and Miscellaneous (MISC).

February 19, 2024

Installation

Clone the repository:

git clone https://github.com/abbaouiAchraf/BERT-NER-COLL2003.git

Install the required dependencies:

pip install -r requirements.txt

Usage

Open the Jupyter Notebook bert-ner-finetuning.ipynb.
Follow the step-by-step instructions provided in the notebook to:
- Load and preprocess the CoNLL-2003 dataset
- Set up the BERT model and tokenizer
- Define the data loaders and the training/evaluation loop
- Fine-tune the BERT model
- Evaluate the model’s performance
- Make predictions on new text data
The notebook includes detailed explanations and code snippets to guide you through the entire process.