ekacare/MedMCQA-Indic

Medmcqa Indic

Dataset Description

This dataset is Indic version of MedMCQA dataset

Languages covered

Language	Code
English	en
Hindi	hi
Tamil	ta
Telugu	te
Kannada	kn
Malayalam	ml
Bengali	bn
Gujarati	gu
Marathi	mr
Punjabi	pa
Odia	or
Assamese	as

This dataset contains 50,185 samples organized across multiple splits and 12 subsets (one for each language).

The translation to 11 Indian languages is done using meta-llama/llama-4-scout-17b-16e-instruct model. Several medical concept names which only exist in English were transliterated (in rare cases left to English, Latin script).

Dataset Structure

Subsets

This dataset includes the following subsets:

• as: 4,183 samples
  • test: 4,183 samples
• bn: 4,183 samples
  • test: 4,183 samples
• en: 4,183 samples
  • test: 4,183 samples
• gu: 4,183 samples
  • test: 4,183 samples
• hi: 4,183 samples
  • test: 4,183 samples
• kn: 4,183 samples
  • test: 4,183 samples
• ml: 4,183 samples
  • test: 4,183 samples
• mr: 4,183 samples
  • test: 4,183 samples
• or: 4,172 samples
  • test: 4,172 samples
• pa: 4,183 samples
  • test: 4,183 samples
• ta: 4,183 samples
  • test: 4,183 samples
• te: 4,183 samples
  • test: 4,183 samples

Usage

Load specific subset and split:

from datasets import load_dataset

# Load specific subset and split
dataset = load_dataset('ekacare/MedMCQA-Indic', 'hi', split='test')

# Load all splits from a subset
dataset = load_dataset('ekacare/MedMCQA-Indic', 'hi')

# Load everything
dataset = load_dataset('ekacare/MedMCQA-Indic')

Dataset Creation

This dataset was created using StreamableDatasetManager on 2025-06-18T22:14:21.768434.

Data Fields

The dataset includes the following columns:

• id: String data
• question: String data
• opa: String data
• opb: String data
• opc: String data
• opd: String data
• cop: Int64 data
• choice_type: String data
• subject_name: String data
• topic_name: String data
• exp: String data
• language: String data

Technical Details

• Total samples: 50,185
• Shard length: 3000
• Number of subsets: 12
• Number of splits: 2