This is a pipeline for performing question-answering tasks using the Stanford Question Answering Dataset (SQuAD) 2.0.
Description This repository contains code to train, validate, and test models for question answering tasks using the SQuAD 2.0 dataset. The SQuAD 2.0 dataset extends SQuAD1.1 to incorporate unanswerable questions, providing a more robust challenge for natural language understanding systems.
The pipeline includes:
- Data loading and preprocessing scripts
- Model architecture scripts
- Training, validation, and testing scripts
- Pretrained models
- Visualization tools
- Evaluation scripts to benchmark your model