bigscience-workshop / biomedical

Tools for curating biomedical training data for large-scale language modeling
454 stars 116 forks source link

Create dataset loader for EHRSQL #879

Open glee4810 opened 1 year ago

glee4810 commented 1 year ago

Adding a Dataset

Name: EHRSQL Description: Large-scale text-to-SQL dataset for question answering (QA) on electronic health records (MIMIC-III and eICU). It covers a wide range of questions asked in the hospital and aims to challenge the trustworthiness of the existing text-to-SQL models by testing whether the model can refuse to answer questions whenever they are not answerable. Task: Text-to-SQL Paper: paper Data: data License: CC BY 4.0