README.md

metadata

language: en
license: mit
tags:
  - question-answering
  - pytorch
  - bert
datasets:
  - rajpurkar/squad_v2
  - Eladio/emrqa-msquad

tiny-clinicalbert-qa

A lightweight, domain-adapted BERT model for clinical question answering, trained on a combination of SQuAD v2 and EMRQA-MSQuAD datasets.

Source code for the training script is available on GitHub. See eval_results.json for evaluation results, and train_results.json for training results.