OctoMed/PneumoniaMNIST

---
dataset_info:
features:
- name: image
dtype: image
- name: image_hash
dtype: string
- name: question
dtype: string
- name: options
sequence: string
- name: answer
dtype: string
- name: responses
sequence: string
splits:
- name: train
- name: test
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
---

# PneumoniaMNIST - Chest X-ray Pneumonia Classification

## Description
This dataset contains pediatric chest X-ray images for binary pneumonia classification. The task involves classifying each X-ray as either normal or indicative of pneumonia based on radiographic features. 16 reasoning traces were collected for each example in this task by sampling with GPT-4o, available in the `responses` column. We greatly appreciate and build from the original data source available at https://medmnist.com

## Data Fields
- `question`: The classification question about pneumonia diagnosis from chest X-ray
- `options`: Multiple choice options representing diagnostic categories
- `answer`: The correct diagnosis
- `image`: Chest X-ray image
- `responses`: Model reasoning responses (in train split)

## Splits
- `train`: Training data with model responses
- `test`: Test data for evaluation

## Usage
```python
from datasets import load_dataset

dataset = load_dataset("OctoMed/PneumoniaMNIST")
```

## Citation

If you find our work helpful, feel free to give us a cite!

```
@article{ossowski2025octomed,
title={OctoMed: Data Recipes for State-of-the-Art Multimodal Medical Reasoning},
author={Ossowski, Timothy and Zhang, Sheng and Liu, Qianchu and Qin, Guanghui and Tan, Reuben and Naumann, Tristan and Hu, Junjie and Poon, Hoifung},
journal={arXiv preprint arXiv:2511.23269},
year={2025}
}
```