Dataset · jopan
Dataset of visual questions and answers from arXiv articles
jopan/FlowLearn
English pairs of images, questions, and answers gathered from arXiv research publications.
What's inside
The repository collects images and English text question-answer pairs for Visual Question Answering (VQA) tasks. Research papers from arXiv served as the source material, and the dataset is distributed under the CC BY-NC-SA 4.0 license. Metadata indicates a small volume (less than 1,000 entries, with the card text mentioning the 1K-10K category).
What you can do with it
- Evaluating and validating Vision-Language Models (VLMs) on scientific illustrations and plots
- Fine-tuning or LoRA adaptation of VQA models for academic papers
- Tasks
- Visual question answering
- Languages
- English
- Size
- up to 1K rows
- Downloads this month
- 36.3K
- Likes
- 8
How to load it
from datasets import load_dataset
ds = load_dataset("jopan/FlowLearn")