Closed
Description
Description:
To help users understand different Question Answering (QA) models, add a notebook that applies multiple methods on the same dataset and compares results.
Tasks:
- Compare traditional rule-based QA (TF-IDF + BM25), extractive models (BERT, RoBERTa), and generative models (T5, GPT-4).
- Evaluate results based on EM (Exact Match), F1-score, and response coherence.
- Summarize key takeaways for different use cases.
- Name the notebook qa_model_comparison.ipynb.
- Update the README file with relevant references.