---
extra_gated_prompt: "Dataset from Shared task on Large-Scale Radiology Report Generation (https://stanford-aimi.github.io/RRG24/). Access requires a verified CITI training certificate using the same process outlined by PhysioNet (see https://physionet.org/about/citi-course/) Please provide proof via the verification URL, which takes the form https://www.citiprogram.org/verify/?XXXXXX. You agree to not use the model to conduct experiments that cause harm to human subjects."
extra_gated_fields:
  Full Name: text
  Email: text
  Affiliation: text
  CITI Certification Verification URL: text
  I agree to the DUA of bimcv-covid19: checkbox
  I agree to the DUA of mimic-cxr: checkbox
  I agree to the DUA of padchest: checkbox
  I agree to the DUA of candid-ptx: checkbox
  I agree to not redistribute or host cheXpert anywhere else: checkbox
dataset_info:
  features:
  - name: source
    dtype: string
  - name: images_path
    sequence: string
  - name: images
    sequence: image
  - name: impression
    dtype: string
  - name: findings
    dtype: string
  splits:
  - name: train
    num_bytes: 72889169962.96
    num_examples: 333205
  - name: validation
    num_bytes: 1885465464.459
    num_examples: 8543
  download_size: 74547230342
  dataset_size: 74774635427.419
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
---


<div align="center">
<h1>
  Interpret-CXR: A Large-scale Collection of CXR datasets.
</h1>
</div>

<p align="center">
📝 <a href="https://arxiv.org/" target="_blank">Paper</a> • 🤗 <a href="https://huggingface.co/datasets/StanfordAIMI/interpret-cxr-public/" target="_blank">Hugging Face</a> • 🧩 <a href="https://github.com/Stanford-AIMI/RRG24" target="_blank">Github</a> • 🪄 <a href="https://stanford-aimi.github.io/RRG24/" target="_blank">Project</a>
</p>

<div align="center">
</div>


## ✨ Latest News
- [02/20/2024]: Shared task at BioNLP@ACL2024 online [[Website]](https://stanford-aimi.github.io/RRG24/).


## 💡 Motivation
We curated the "Interpret-CXR" dataset for the following motivations:
* For the shared task on large-scale radiology report generation at BioNLP@ACL2024.
* Simplify the data access process.
* Standardize the benchmark for future research in this field


## 🎬 Get Started
```python
from datasets import load_dataset

dataset = load_dataset("StanfordAIMI/interpret-cxr-public")
```

## ✏️ Citation

```
TBA
```