Fine-tuning Question-Answering Model for COVID-QA dataset

Full-finetuning Question-Answering model using the COVID-QA dataset related to COVID-19.

Dataset

The dataset used for this project is the COVID-QA Question Answering Dataset.

Project Structure

src/
├── config/
│   └── cfg.yaml
├── data/
│   └── dataset.py
├── models/
│   └── model.py
├── scripts/
│   ├── test.py
│   └── train.py
├── utils/
│   └── utils.py
└── app.py
requirements.txt

Model Information

RoBERTa (Robustly optimized BERT approach) - roberta-base

Fine-tuning the Model

To fine-tune the model, follow these steps:

1. Install Dependencies: Install the required dependencies using the following command:

   pip install -r requirements.txt

2. Prepare the Dataset: The dataset is automatically loaded and preprocessed by the scripts.

3. Train the Model: Run the training script to fine-tune the model:

   python src/scripts/train.py

4. Evaluate the Model: Run the testing script to evaluate the model:

   python src/scripts/test.py

5. Run the App: Run the Streamlit app for real-time question-answering with the model:

   streamlit run src/app.py

The configuration for the model, training, and evaluation is specified in the cfg.yaml file. Adjust the configuration as needed for your specific requirements.