metadata

language: fa
pretty_name: Clinical Question Answering Dataset I (Farsi)
tags:
  - Farsi
  - Persian
  - clinic
  - QA
  - question
  - answer
license: cc

Clinical Question Answering Dataset I (Farsi)

This dataset contains approximately 50k questions and around 60k answers, all produced in written form. The questions were posed by ordinary Persian speakers (Iranians), and the responses were provided by doctors from various specialties.

Dataset Description

Question records without corresponding answers have been excluded from the dataset. This dataset is NOT a part of Clinical Question Answering II dataset and is a complete different dataset.

Usage

Huggingface datasets library:

from datasets import load_dataset
dataset = load_dataset('PerSets/cqai')

Pandas library:

import pandas
df = pd.read_json("train.jsonl", lines=True)

Vanilla Python:

import json

train = []
with open('train.jsonl', encoding="utf-8") as f:
    for line in f:
        obj = json.loads(line)
        train.append(obj)

License

CC0