/Question-Answering-CovidQA

Questions answering (zeroshot, finetuned, adapter model) on covid-qa

Primary LanguagePython

Out of domain Question Answering

This repo contains the code for out-of-domain question answering on covid question answering dataset using different models. Huggingface library is used to run all models on deepset/roberta-base-squad2 model. The following models are used:

  1. Zero-shot inference: In order to build baseline for further models, zero-shot inference using HF question-answring pipeline is done. This gave an Exact Match of 24.2 on Test set. Run roberta_zeroshot.py for inference.

  2. Finetuning: The model is finetuned on covid dataset for 3 epochs. This gave Exact Match of 35.4 on Test set. Run roberta_finetune.py for training and inference.

  3. Adapter Model: Instead of finetuning complete model, adapter layers are used which reduce the learnable parameters by a large amount (130X). See AdapterHub for more information. When trained on 3 epochs, the exact match on test is 33.8. Run roberta_adapter.py for training and inference.