Skip to content

Dataset · derek-thomas

ScienceQA: Scientific questions with images

derek-thomas/ScienceQA

A multimodal English science question dataset with multiple-choice options and images.

What's inside

ScienceQA is an English multiple-choice science question dataset. Examples include an image, a question, and answer choices; the card describes it as multimodal reasoning with chains of thought. Size is 10K-100K.

What you can do with it

  • Training a VQA model for science questions
  • Fine-tuning LoRA for multimodal reasoning
  • Evaluating models on a multiple-choice benchmark
Tasks
Multiple choice, Question answering, Other, Visual question answering, Text classification
Languages
English
Size
10K–100K rows
Downloads this month
31.3K
Likes
244

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("derek-thomas/ScienceQA")