metadata

tags:
  - distilbert
  - health
  - tweet
datasets:
  - custom-phm-tweets
metrics:
  - accuracy
base_model: distilbert-base-uncased
model-index:
  - name: distilbert-phmtweets-sutd
    results:
      - task:
          type: text-classification
          name: Text Classification
        dataset:
          name: custom-phm-tweets
          type: labelled
        metrics:
          - type: accuracy
            value: 0.877
            name: Accuracy

distilbert-phmtweets-sutd

This model is a fine-tuned version of distilbert-base-uncased for text classification to identify public health events through tweets. The project was based on an Emory University Study on Detection of Personal Health Mentions in Social Media paper, that worked with this custom dataset.

It achieves the following results on the evaluation set:

Accuracy: 0.877

Usage

from transformers import AutoTokenizer, AutoModelForSequenceClassification
tokenizer = AutoTokenizer.from_pretrained("dibsondivya/distilbert-phmtweets-sutd")
model = AutoModelForSequenceClassification.from_pretrained("dibsondivya/distilbert-phmtweets-sutd")

Model Evaluation Results

With Validation Set

Accuracy: 0.8708661417322835

With Test Set

Accuracy: 0.8772961058045555

Reference for distilbert-base-uncased Model

@article{Sanh2019DistilBERTAD,
  title={DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter},
  author={Victor Sanh and Lysandre Debut and Julien Chaumond and Thomas Wolf},
  journal={ArXiv},
  year={2019},
  volume={abs/1910.01108}
}