Dataset · flaviagiammarino
PathVQA: pathology visual question answering
flaviagiammarino/path-vqa
Question-answer pairs on pathology images for training and evaluating medical VQA systems.
What's inside
PathVQA contains question-answer pairs on pathology images. Questions are open-ended or binary yes/no. The dataset was compiled from two public pathology textbooks and a public digital library. It is intended for training and testing medical VQA systems.
What you can do with it
- Train a medical VQA model
- Evaluate VLMs on pathology images
- Fine-tune LoRA for pathology question answering
- Tasks
- Visual question answering
- Languages
- English
- Size
- 10K–100K rows
- Downloads this month
- 22.5K
- Likes
- 77
Models trained on it
- sarathi-balakrishnan/Qwen2.5-VL-7B-Medical-LoRAImage text → text 43
- Reverb/Idefics2-8b-docVQA-finetunedImage text → text 22
- marwanimroz18/blip-trainerVisual question answering 18
- trieutm/blip-vqa-finetunedVisual question answering 14
- dhairyapandya/pathos-gemma4-distilled-rl-GGUF 12
- usama10/paligemma2-3b-pathvqaImage text → text 0
How to load it
from datasets import load_dataset
ds = load_dataset("flaviagiammarino/path-vqa")