README.md

metadata

license: mit
language:
  - en
base_model:
  - distilbert/distilbert-base-uncased
pipeline_tag: text-classification
tags:
  - coronanet
  - BERT
  - DistilBERT

coronaBERT: DistilBERT-Based Model for COVID-19 Policy classification

Overview

• This repository contains code, training scripts, and documentation for a fine-tuned DistilBERT model for national policy related to COVID-19 classification tasks. The model is designed to classify text descriptions of COVID policies into predefined categories (Policy Type). The project includes data preprocessing, model training, evaluation, and quantization for efficient inference.

Feature

• Fine-tuned on custom labeled data for sequence classification
• Supports dynamic quantization for reduced model size and faster inference
• End-to-end pipeline: data preparation, training, evaluation, and model export
• Ready-to-use with Hugging Face Transformers

Model detail

• Base model: distilbert-base-uncased
• Data: CoronaNet Research Project
• Task: Sequence Classification (e.g., predicting type)
• Framework: PyTorch, Transformers
• Quantization: Optional INT8 dynamic quantization for deployment

Citation

@misc{joesh1-coronaBERT,
    author = {Joesh1},
    title = {coronaBERT: DistilBERT-Based Model for COVID-19 Policy Classification},
    howpublished = {\url{https://huggingface.co/Joesh1/coronaBERT}},
    year = {2025},
    note = {Accessed: [Insert Date Accessed, e.g., 2025-05-22]}
}